Part of a series of occasional features marking the 250th anniversary of the Declaration of Independence

Celebrations of the 250th anniversary of the Declaration of Independence have called attention once again to the oft-repeated origin tale of how the revolutionary document sprang from Enlightenment ideas — natural rights, the social contract, and separation of powers.

But other parts of the story often get less attention. That is how the founders’ thinking was affected by existing social and political conditions on the ground, as well as their familiarity with the evolution of English constitutional history.

The independence movement really took hold when political leaders grew angry over the imposition of Parliament’s tax laws, such as the 1764 Sugar Act, the 1765 Stamp Act, and the 1767 Townshend Acts that imposed tariffs on British goods. And their efforts to oppose British rule were bolstered by early forms of self-government that had emerged in the American colonies. 

“Town meetings and colonial assemblies made the Revolution happen,” said Robert Allison, Ph.D. ’92, a Suffolk University historian who studies early American colonial history. “Parliament declared that it had the power to govern the colonies ‘in all cases whatsoever,’ but each colony already had an elected assembly which made laws for it. And in New England each town had a meeting — at least once a year, but in many cases when any 10 tax-paying adult men signed a warrant for one — that made laws and elected all officials in the town.”

As the movement took shape, the framers of the Declaration were heavily influenced by 17th century political philosophy, specifically John Locke’s ideas of natural rights and government by consent and Montesquieu’s doctrine of separation of powers.

But they had to adapt some those ideas, many of which had never been tried in practice, according to Michael Klarman, Charles Warren Professor of American Legal History at Harvard Law School.

One example is the way they took Locke’s foundational formula of universal natural rights, “life, liberty, and property,” and substituted the right to property with “the pursuit of happiness” in the Declaration of Independence.

The antecedents to the phrase “the pursuit of happiness” can be traced to James Wilson, a signer of the Declaration and the Constitution, said Danielle Allen, James Bryant Conant University Professor.

Wilson is one of the least-known founders, but a recent book, “The Lost Founder: James Wilson and The Forgotten Fight for a People’s Constitution” by journalist Jesse Wegman, sheds more light on his role in the country’s founding, Allen said.

Essentially, Wilson advanced the idea that one of the central purposes of government should be something broader and higher than the preservation of individual property.

“Wilson drew on the Scottish Enlightenment for a common-sense approach to human morality that put happiness at the center, as a concept of human flourishing and well-being,” according to political philosopher Allen. “And then John Adams picked up that line of argument and used the vocabulary of happiness to make the case for what kinds of government should be formed throughout the fall of 1775 and the spring of 1776.”

Also by the spring of 1776, the right to property had been connected to a defense of enslavement and had become a contested term, she said. “The Lockean triad was life, liberty, and property. So, life, liberty and the pursuit of happiness was keeping the issue of slavery out of that important preamble.”

As well-educated people, and as British subjects themselves, the founders were well-versed in 17th and early 18th century English history, a turbulent period during which England became a constitutional monarchy, in which the king could no longer act without Parliament’s support.

“This was part of the cultural formation of the people who made the revolution,” said Eric Nelson, Robert M. Beren Professor of Government. “These were people who considered themselves as Englishmen, and they were familiar with the entire tradition of English constitutional history, including the fearsome debates over liberty, sovereignty, monarchy, and property going back at that point nearly 600 years.”

The founders drew upon many Enlightenment intellectuals, including Locke and his influential “Two Treatises of Government” (1690). But Locke was only one of many thinkers who influenced the founders, said Nelson.

“Locke was a major figure within the traditional of political thought and constitutional refection that the founders were familiar with,” said Nelson. “He was a very important source, but it’s wrong to say that it was ‘The John Locke Show.’”

Among other political theorists of the era was James Harrington, who wrote “The Commonwealth of Oceana” (1656) on what makes an ideal republic and what makes a government stable.

Also influential were Algernon Sidney, author of “Discourses Concerning Government” (1698), a critique of royal absolutism and a defense of republicanism and popular government and the natural equality of all men; John Trenchard and Thomas Gordon, who wrote “Cato’s Letters” (1720-1723) on the importance of liberty, government limits, and the rule of law; and David Hume, who penned “A Treatise of Human Nature” (1739-1740) on liberty and checks on powers.

In addition, the 1215 Magna Carta, which set limits on royal power, and the English Bill of Rights of 1689, which limited the powers of the monarchy and established Parliament’s rule over the British Crown, were part of the intellectual origins of the founders’ vision for the new republic.

“The most important intellectual source they had was an understanding of 17th and early 18th century England, when there had been a big fight over the role of Parliament versus the role of the Crown, and they were really versed in that whole political world as though it was yesterday,” said Allison.

Even as the ideas of liberty and self-governance drew popular support to the independence movement in the colonies and abroad, there was considerable disagreement among the founders about the new kind of political system they were building.

“We spend a lot of time in our political culture asking, ‘What did the founders think about X?’” said Nelson, the government professor. “The founders were a very heterogeneous group of people who thought very different things and had very different political and constitutional preferences and ideas. So, the question ‘What did the founders think about X?’ is almost always badly posed.”

The result was a lot of compromise, improvisation, and much experimentation.

For that reason, Klarman, who wrote “The Framers’ Coup: The Making of the United States Constitution,” warns of the dangers of treating the nation’s foundational documents as “divinely inspired” and its framers as “demigods.”

“I think the framers were pretty virtuous, and they were trying to do what was good for the country, but they had interests, just like everybody has interests, and among them was to protect property rights,” said Klarman. “But we turn the founders into God-like creatures, who were ultra-virtuous statesmen who had no interests or prejudices, but simply were divining the best form of government.”

It’s an argument that prevents necessary changes in the Constitution, he said, to make institutions more responsive to these modern times and the union more perfect.

“Throughout American history, there has always been this tendency, usually a conservative-leaning tendency, to say, ‘How dare you touch the Constitution? How dare you criticize how the Senate is constructed? How dare you criticize the Electoral College?’ and that we ought to be reluctant to change any of these things because the framers were so extraordinary, so selfless, so divinely inspired … We should decide these questions on their merits, not with a heavy thumb on the scales in favor of what the founding generation believed.”

More like this

• 

  Hot off the presses: The Declaration of Independence

  A flawed first copy — dashed off in all-nighter — endures as template for democracy in action

• 

  America the beautiful

  Celebrating nation’s natural wonders through art on 250th anniversary

• 

  One of first to receive University honorary degree? George Washington.

  Archivist says it’s an early example of ‘long tradition of honoring national leaders’

• 

  How did Ben Franklin’s barber keep up with news on war for independence?

  Houghton exhibit showcases newspapers, pamphlets, other media, offering on-the-ground view of what regular citizens knew, and when they knew it 

• 

  Helping to give birth to nation — and Harvard Med

  School founder John Warren numbered among alumni who were part of revolutionary generation

• 

  Walking in Harvard’s ‘Revolutionary footsteps’

  Exhibit traces University’s role in America’s birth — from campus barracks to Founding Father alumni

• 

  Did the British unleash biological warfare against Washington’s troops?

  Historians trace role of physicians, medicine, disease during war in articles marking 250th anniversary of Declaration of Independence

• 

  High-tech for the 1700s

  Exhibit showcases tools that powered Revolutionary America

Part of a series of occasional features marking the 250th anniversary of the Declaration of Independence.

The U.S. boasts a diverse array of stunning natural landscapes, from the dense Louisiana bayous to the peaks of the Rockies. 

“Nature is our commonwealth — our common wealth,” said Horace D. Ballard, Theodore E. Stebbens Jr. Curator of American Art at the Harvard Art Museums.

Early European colonists in the Americas, Ballard explained, sent friends and colleagues plants, seeds, and the bones of mastodons, as evidence of the marvelous landscapes that produced them. Later, in the 1850s and ’60s, Americans with little ability to travel would pay to crowd into tight rooms for a chance to view a single painting of Niagara Falls. 

“When it comes to the United States, with all of its thousands of microclimates and ecosystems, it’s dizzying,” Ballard said. “One of the most profound and simple — but not naive — pronouncements we can make and hold onto on the country’s 250th anniversary is this incredible, dizzying diversity we’re all so proud of, and we all benefit from, socially and environmentally. E pluribus unum — out of many, one — defines the way the world has always seen the U.S.”

You can find many examples of artwork documenting the nation’s natural beauty at Harvard Art Museums. Ballard and other experts walk us through a selection below. 

---

---

---

---

---

---

---

---

It has become a science fiction trope come to life.

Tech giants OpenAI and Anthropic disclosed in July that versions of their AI models broke through safeguards known as a “sandbox, gained access to the internet, and hacked servers of outside companies during internal tests of their cybersecurity capabilities.

This week, AI Security Institute, a U.K. government research entity reported additional incidents, saying AI agents created fake online personas in order to improperly access real people and companies during security tests it conducted on the two tech firms.

OpenAI CEO Sam Altman characterized the July breach as an “unprecedented” and “significant security incident” caused by rogue AI agents. Its investigation has since uncovered evidence of additional breakouts. Anthropic blamed human error involving an evaluation partner.

In this edited conversation, James Mickens, Gordon McKay Professor of Computer Science at SEAS and director of the Berkman Klein Center at Harvard, explains what these breaches mean for the future of AI security.

---

Gazette: How plausible are the explanations offered by OpenAI and Anthropic?

James Mickens: It’s hard to say. I think it is very commendable that both OpenAI and Anthropic decided to release public incident reports about what happened. They could have very easily kept this to themselves and not let people know that these models have these autonomous abilities to get past these sandboxes.

However, because the public doesn’t really know the details of what happened, it’s hard to fully verify the timeline and the narrative.

The explanations put out by OpenAI and Anthropic are certainly plausible. Those public-facing stories are certainly what could have happened. Is that literally what happened? Was there anything left out? We don’t know.

We do know both companies have talked to third-party security companies to validate stuff. But we don’t really know what’s going on.

And this points to one of the tensions with this dangerous technology. One can understand why OpenAI might not want to say, “Here’s a detailed description of our sandboxing environment.”

But on the other hand, there’s this very well-known security principle, which is that sunlight disinfects. The more eyes you have on something, the more brains you have to try to find problems.

Two of the biggest players in AI report similar incidents within weeks of each other. Is that troubling?

Yes, it’s quite concerning. I would also say these are not threats that the security community and the AI safety community are just realizing were possible.

Here at Harvard, we’ve been saying for a long time that security researchers need to be thinking more critically about strengthening sandbox mechanisms and making it more difficult for models to break out of them. Last year, my research group published a paper specifically about how to create new types of software and hardware to make sandbox escapes more difficult.

Even as recently as last year, there were some people saying, “You’re guarding against a sci-fi eventuality. You’re worried about things that really are not top of mind for people using these AI systems in the real world.”

Yes, there are frontline risks that are harmful and relatively mundane, but then there are other risks that are exceptionally pernicious and quite alarming in terms of how quickly they can affect huge numbers of people in society.

For example, what would happen if one of these models tried to break into the power grid or tried to tamper with financial systems to affect the stock market? These are very real concerns.

I think with what we’ve seen over the past week or so, we should be worried about this problem more generally.

These models are very difficult to understand. There’s a whole field of AI interpretability, which seeks to understand why models do the things that they do. They’ve made some progress, but it’s still an open challenge to guarantee that a model will always act in ways that are aligned with human preferences.

So yes, these incidents are quite concerning.

Is that a concern about the technology or about the people overseeing the technology and their priorities?

Both. First and foremost, there are technical challenges in defining what alignment means, determining whether models are acting in aligned ways, and then enforcing human-defined alignment policies when models misbehave.

Those challenges are technical inasmuch as they require us to build new software and new hardware to enforce alignment policies.

However, there are also social problems of governance. Who are the humans who decide what misalignment and alignment mean? Are they mostly people who live in the U.S.? Who work for U.S. companies? Are they people who live in China, or in India, or in Brazil? Are these questions that should be answered by companies or should governments be doing this? And how do we ensure international coherence of approaches toward these things?

So, there are some very thorny technical and socio-technical issues here that are intrinsically difficult to solve and that require thoughtful responses that are at odds with the breakneck speed that companies want to move at.

Do these incidents suggest that AI models have now moved beyond human control?

I would say these incidents demonstrate that even frontier lab companies cannot guarantee model alignment in all scenarios. We’ve actually known for some time that defining aligned behavior is difficult, enforcing aligned behavior is difficult, and that models are already sufficiently complicated and sophisticated to press us on those two issues.

In other words, AI safety risks are not a theoretical problem. They haven’t been for quite some time. And it’s naive for people to think that we can afford to not think about safety risks because we want these AIs to solve cancer or things like this. We have safety problems now. We can’t wait to think about this.

Given the winner-take-all stakes in AI, could such unflattering disclosures be a kind of humblebrag by OpenAI and Anthropic to show how cutting-edge their AI models are?

Yes, it’s certainly possible.

This is why a lot of security researchers, a healthy minority, take all of these releases with a grain of salt because first of all, the public has no way of knowing whether these types of incidents have happened 100 times this year. And we’re just hearing about the recent ones now because why?

Could it be that all the companies want to race toward Artificial General Intelligence [AI capability that matches or exceeds human-level skills] and claim the AGI mantle first? AGI is less a crisply defined technical achievement and more of a “we’ve decided that we have it and now we’re going to claim it.”

And so, cynics in the security community are saying, “We bet that sandbox escapes have happened before. You just didn’t tell us. The reason you’re telling us about these sandbox escapes now is you’re trying to pave the way for this declaration of AGI unilaterally, and you’ll be able to point to this press release and say, ‘See? That was the first glimmering of our model which has AGI.’”

There’s a lot of stuff these companies put out that technical folks and policy folks don’t take fully at face value.

Is it too late to impose regulations on AI to try to prevent this from happening again?

It’s not too late to impose regulations.

Getting the right regulations in place will be tricky because this is a socio-technical problem that requires both robust technical mechanisms for sandboxing models and societal consensus on what the trade-off should be between speed of AI development and reining in potential harmful behaviors from these models. We don’t have that second part either.

Different reasonable people will have different ideas of how we should be striking those trade-offs. The important thing, though, is that we at least have to have the conversation seriously. That’s, I think, the big thing.

What types of interventions might be most effective?

There’s been a variety of mechanisms, technical, legal, and otherwise, that have been proposed for making models safer.

I think a key challenge, though, is how do we ensure that the model vendors are actually using these mechanisms? How do we get more insight into what happens when something goes wrong? How do we understand who should be held accountable for these things?

It’s important for us to not be unnecessarily restrictive. We do want to allow these companies to engage in cutting-edge research. But, at this point, the risks are getting so large and so unavoidable that it would be irresponsible for us as a society to not think about how to reckon with these things.

It’s not as if we have no levers that we can push to try to make safety better. There are a lot of people who are thinking about how do we interpret what models are doing, how do we steer them to act in ways that are more societally aligned, and so on.

There are also regulators and people in government thinking about this too, so it’s not as if there aren’t people who care about AI safety risks.

Hopefully, this past week can be a fulcrum moment where we have a series of incidents that are so clear in terms of their implication that society will have to organize around this and try to think critically about AI safety.

Excerpted from “Gerontocracy in America: How the Old Are Hoarding Power and Wealth ― and What to Do About It” by Samuel Moyn, J.D. ’01, and published by Farrar, Straus and Giroux

For two centuries, “Democracy in America,” published by the Frenchman Alexis de Tocqueville starting in 1835, has been the most famous book about the United States. It deserves to stay that way. But it might need an update.

There have always been older people. A psalm of the Bible says that, unless they die early for some reason, human beings get 70 years — 80 if they’re lucky. Genesis mentions generations of superagers at the start of humanity — like Adam, Noah, and Methuselah, the latter of whom celebrated his 969th birthday. And elders have long been accorded preferences and priorities. Whenever seniority is used to allocate positions or resources, an impulse toward gerontocracy — the empowerment of the old — is at work. Another biblical passage, this one from Proverbs, holds that “gray hair is a crown of glory.” The “preference for age and seniority,” one historian has remarked, has been “shared by all those corporate institutions which set a value on hierarchy, stability, and continuity.”

But there is something new that has supercharged elder power. Starting in Tocqueville’s 19th century, a “great aging” took place as average life expectancies increased, thanks to the reduction of childhood mortality and eventually adult disease. Today, a new crop of Methuselahs is living beyond 110 or even 120, in the case of Frenchwoman Jeanne Calment, who died in 1997 at 122, the oldest age on record. But the real change is not that more people live to 100 but that so many reach 80 or 90, even as fewer are born. The political effects of a citizenry with a higher median age have been mounting.

At the dawn of modernity, Tocqueville commented that a new political science was needed for a new age; in the 21st century, an even newer political science is needed for the new aging. 

The virtue of life extension, obviously, is that life is generally preferable to death — especially if the debilitating effects of age are also driven into the future. I remember my father, as he progressed toward the unavoidable, often counseling me never to get old; one of my mother’s nuggets of wisdom, as she has lived past 80, is that the alternative to aging is even worse. “Death, no matter how inevitable, is an abrupt cancellation of indefinitely extensive possible goods,” the philosopher Thomas Nagel writes, adding that death is no less tragic for those lucky enough to live long rather than short lives. 

But there are vices to a culture of extended life — and one of the biggest is political, in the broadest sense. An aging society is more set on preservation than on renovation. And then there’s the unfairness involved in blocking the young and hoarding their inheritance while mandating them to care for the old and requiring them to serve a government helmed by their elders.

Americans definitely worry today that our democracy is at risk — that new forms of oligarchy are a frightening possibility. But what if we have long been drifting into an unsuspected form of rule, one less apparent because it is only sometimes visible in government: a new kind of society that privileges many older Americans — to the detriment of all, including many senior citizens themselves?

Welcome to gerontocracy in America.

←→

President Donald Trump is old. His predecessor, Joe Biden, is old, too — and was even older-seeming when it counted. The evidence of Biden’s decline during the 2024 election campaign altered the course of American history, leaving a storied republic on the brink.

For years, loud voices had cautioned against even broaching the topic of Biden’s debilitation. One observer went so far as to boast that gerontocracy serves America well. Even he flip-flopped in the face of Biden’s undeniable limits. The experience of 2024 brought the political consequences of aging leadership home to millions of American citizens. Of course, Democrats continued to show a default respect for the aged — except for when they were combating Trump by highlighting his own age-related dysfunction.

Of course, not all aging politicians are decrepit. Nor is the cognitive and physical decline of politicians — a genuine problem beyond the presidency, with the median age of congresspeople spiking to over 60 and of federal judges to 68 — the main effect of aging on politics.

An aging society is more set on preservation than on renovation.

Our political system is democratic, meaning that the authorities — no matter their age — depend on electoral and other popular support to rule. And our society is capitalist, so most actual power is exercised by people outside government, especially the wealthy. To inveigh merely about the age of politicians is, in many ways, to miss the point of how power works today. If the voters and the wealthy are old, it may not even matter whether the politicians are young.

To a gobsmacking extent, seniors, with their high rates of participation, dominate elections, especially the apparently boring local and off-cycle ones. The median age of actual voters these days — meaning that equal numbers of voters are older and younger — is about 52, and on the way up. Outside presidential contests, it is 55 or 56; in primaries, it is 65. Even older funders disproportionately finance the races. The most common age of donors in recent elections can run as high as 70. Finally, older Americans are backed by the best-funded advocacy group in world history — the AARP — and enjoy the most governmental benefits by a vast degree. The government is bought and paid for by those who are up in years and is organized for their sake. Trump’s “big beautiful bill” is only the most recent exacerbation, cutting down the welfare state — except elder entitlements — to make up for lost taxes. 

It doesn’t stop there. Aging Americans control the biggest bank accounts and stock portfolios, partly as a function of living long enough to accumulate more and more wealth without giving much away. In our time, the old rich have pulled away from everyone else. In the two decades or so straddling 2000, households headed by adults older than 65 improved their median net worth by 42 percent, while the wealth of families of adults 18 to 34 fell by 68 percent. American oligarchy is, by and large, an “oldigarchy.” 

Older people are also hoarding jobs. Following the abolition of mandatory retirement in most professions, older people are choosing to stay in some of the best positions, accumulating more money and blocking pipelines. The average age of CEOs and professors alike has gone through the roof. The United States, which has the highest wage inequality by age in the world, has seen the numbers gallop: Between 1979 and 2018, the pay gap between those over 55 and under 35 has increased by 61 percent. In those same years, the proportion of older people in the workforce has also exploded: The share of workers over 55 rose a shocking 88 percent.

The elderly possess the choicest real estate in great cities. The housing crisis, now well-known, is partly a story of old people refusing to downsize as they age. The age group most likely to own a home in America, at a rate of over 80 percent, is 70 to 74, with those 75 and older second highest. Shockingly, the age of the median homebuyer leaped from barely 30 in 1981 to 53 in 2022. The failure of abundant supply is also partly the fault of the elderly, since their political mobilization arrests development and blocks new builds. In defense of their incomes, wealth, homes, or land, they have driven tax revolts so that they don’t have to contribute to the common good.

In short, it’s not just that presidents and other politicians are old. It’s not just the cognitive or bodily decline they suffer. Rather, what’s most important is that such leaders represent an aging constituency that controls the political system. They are also the visible face of the elder domination of private forms of power, especially wealth. What’s more, with millions privileged by gerontocracy, there is no way to separate the age of our elites from their ascendancy. In America today, age is the modality in which class is lived (with apologies to the great, late cultural theorist Stuart Hall, who said the same thing about race).

Today’s gerontocracy is not anybody’s malevolent plan, exactly. It is more an accidental by-product of a legitimate and understandable desire to survive, with political effects that nobody has reckoned with. Another day, month, or year among loved ones — what downside could that have? Such a question would have been rhetorical until recently, but it has now snuck up on unsuspecting societies. Boomers aren’t distinctively evil (well, some are). Rather, the fact that they are so numerous, and the fact that they are aging in an era when the extension of life is normalized, makes an endemic syndrome more obvious.

That syndrome is a gerontocratic crisis of succession on the scale of American society itself. It often involves a deathwatch for old people. The melodrama of succession — waiting for the old to make way for the new and give up effective power — is a feature not only of our politics but also of our economy and of our culture writ large.

Should Trump fail, for whatever reason, to survive beyond the midpoint of his second term, J.D. Vance would succeed him as the youngest president in all American history, beating Theodore Roosevelt, who occupied the office just shy of his 43rd birthday. Even then, American gerontocracy would remain. It would retain its grip on the rest of government, and on our society as a whole.

←→

If Americans deserve a society that is oriented to innovation and problem-solving and equips and launches the young into the prime of life, the country is heading in the wrong direction. “They live in memory rather than hope,” the Greek philosopher Aristotle observed in his ancient diatribe against the aging and in defense of people in their prime.

The presumption that people grow more conservative as they age is one of the hoariest political clichés — but it is nonetheless true. Of course, these days the far right often draws from the very young. Even so, it’s abundantly clear that age and outlook tend to correlate. Who believes the children are our future? Not American seniors. The politics of educational funding for public primary and secondary schools shows a definite connection between aging and stinginess — with voters beyond their child-raising years preferring lower taxes to decent schools. Far beyond education, elder priorities, according to experts, have come “at the expense of investments” that “promote long-term growth” such as housing and infrastructure. Enterprise in the broadest sense is not a priority of gerontocratic societies. And it has been demonstrated that older voters register more sensitivity to inflationary pressure on the continuity of their way of life than to low growth, which they do not feel affects them.

Shockingly, the age of the median homebuyer leaped from barely 30 in 1981 to 53 in 2022.

The oldest are also the most implacable opponents of immigration — even though newcomers are the likeliest to provide the care an aging American population needs. One reason for the disconnect is that older Americans incarnate the demography of the past. Ethnic and racial diversity in the United States is at its maximum in current infants and lowers all the way up the age continuum to the 85 to 99 set. The most common age for whites is 58 — more than double the most frequent age for minorities. This wide disparity, partly due to existing immigration, sets up a choice between preserving older people’s conception of the America they knew and allowing immigrants to fulfill the basic needs of older people themselves.

Most important, gerontocracies are prone to let long-term problems fester and worsen. As the individual life dwindles, playing for time in the face of impending catastrophe is a psychologically appealing stratagem of avoidance and denial. The aged preference for addressing the short term leads to the imposition of heavier and heavier consequences for the future — as those understandably concerned about an indebted nation or a warming planet stress.

A fundamental imperative is therefore at stake. Americans must transcend gerontocracy if we are to realize our collective aspiration to innovation, as well as achieve other important goals such as intergenerational equity, fair political representation, and even an existential acceptance of limits as we age. Everyone likes to think of themselves as ageless, but we are finite, with just one passing life to live. Ultimately, gerontocracy is built on a denial of this transience. The aging often respond to mortality by holding on to power as long as they can — but if life is inevitably lived in stages, then succession doesn’t need to be as melodramatic or unfair as it is now.

←→

If modern gerontocracy functions through a new political environment, it follows that a new range of solutions is required. Age limits for political office are a must. Other fixes range from amplifying the political voice of younger voters and disrupting the stranglehold of the old people’s lobby on our politics to policies that facilitate intergenerational wealth transfer. The schemes I offer seek to advance “intergenerational equity.” But unlike the most vocal partisans of the cause, I don’t oppose the American entitlement state for seniors. Such voices insist that an end or reduction of perquisites like Social Security must occur in the name of fairness. The truth is the opposite: only an expansion of our generosity, including the promise of long-term care, can overcome the iron grip of old people on power.

While gerontocracy is hardly the only source of our difficulties, no one interested in how injustice and social strains are shaped and reproduced in America today should ignore how deep it goes. Gerontocracy has to be assessed properly, for the sake of a new intergenerational political compact. In the end, taking it seriously is essential for treating everyone appropriately and fairly.

There is still a chance for a reset. The extension of the lifespan has had political consequences, many for the worse. It has already brought an ancient phenomenon — the authority of elders over everyone else — back in a dangerous new guise. Biden and Trump have exposed one part of our gerontocracy. Ripping away the curtain that hides the rest can prepare us to dismantle the system and create something new.

Copyright © 2026 by Samuel Moyn. All rights reserved.

George Usmanov did some unforgettable sightseeing this summer.

Over 25 days in June and July, he passed by ghost towns from both the 19th and 20th centuries and acres of forest, “completely charred.” He saw a river run yellow out of a Superfund site and the sun rise purple, thanks to a mix of “desert dust, mist from the Rockies, smoke that was coming in from wildfires.”

Typically, though, he had no one with whom to share those moments. Usmanov was taking part in the Tour Divide, widely regarded as the world’s most difficult bike race. It stretches over 2,700 mostly unpaved miles from Banff in southern Alberta, Canada, to the U.S.-Mexico border. And during his biking days, “about 70 percent of the time, I was entirely alone,” he said. “Even if you’re only separated from other racers by a couple of miles, the trails are extremely twisty and you don’t see them.”

This was in part a working holiday for Usmanov — and that kind of isolation was his object of investigation.

The Harvard Medical School research fellow in neurology at Massachusetts General Hospital is engaged in postdoctoral research on stroke — the nation’s fourth-most-common cause of death, and a condition in which every minute counts. The distances he covered are the same ones that separate some of the continental U.S.’s most isolated communities from centers of effective treatment.

Usmanov, 28, is not a doctor but a medical sociologist. His studies focus on “illness experiences”: how felt symptoms, social connections, and cultural norms encounter and shape treatment.

Stroke — with its various forms and subtle symptoms — is of obvious interest. Most stroke patients don’t feel pain. Instead, at onset, they experience “numbness, tingling, or some sense of disruption: you’re driving a car and suddenly you can’t turn the steering wheel. Or you’re using the TV remote and suddenly you don’t know how to change the channel,” Usmanov said.

Meanwhile, though, a clock has started ticking. “From that moment, parts of your brain are dying, at a rate of around 1.9 million neurons a minute,” Usmanov said.

In cases of acute stroke, the field of neurology has come to see the first 60 minutes after onset as the “golden hour”: If patients are examined and treated with an intravenous clot-busting drug like alteplase within that window, their odds of a full recovery improve dramatically. And clot-busting drugs can be used effectively for up to 4½ hours after a stroke’s onset.

The trouble, Usmanov said, is that “in the United States, the vast majority of people don’t get to a hospital in time, meaning they’re missing out on the most effective treatment.” (Mechanical thrombectomy — using a stent or vacuum to remove a clot physically — can be used up to 24 hours after onset, but it’s more invasive.)

To be sure, even most acute strokes are not lethal. But up to one-quarter of acute stroke survivors die within a year of onset, and survival and recovery rates vary greatly by stroke type, patient age, and the speed and type of treatment. In general, Usmanov endorses what has become a neurological maxim: that in cases of stroke, “time is brain” — meaning there’s an immense public health interest in shortening the gap between onset and appropriate treatment.

Usmanov noted that this country has made important strides toward making “golden hour” arrivals more likely — at least in theory. Individual medical centers have worked to shorten “door-to-needle” time: from intake to radiology and infusion. And not only can ambulances or helicopters shorten trips from homes to “stroke-capable hospitals,” there are more of those than ever before.

A team of Harvard researchers found that in 2019, 91 percent of Americans lived within one hour (by ambulance) of a center that could deliver alteplase; in 2011, it was just 80 percent. Further, they found an additional five percent can be reached by “telestroke”: a remote examination over video that has proven effective in lowering mortality and improving outcomes.

But one large region has remained stubbornly above the national average — the Mountain West, the “exact corridor” through which the Tour Divide passes, Usmanov said.

Usmanov has been an avid biker since his youth in Minneapolis. And this isn’t the first time he’s used his bike as a platform for rolling ethnography. During his graduate studies at Georgia State University, he took a similar ride to visit the informal communities that arise at campsites and RV parks across the Deep South.

“A bike is just an extremely useful and effective way to witness space, I think,” he said from his office in Somerville.

But in the Rockies, he found community harder to come by. There are fundamental problems of logistics — long distances, snowed-over mountain passes closed each year, unmaintained roads even in summer. Isolation is the rule even when ties are close. “The cost of seeing each other is so high,” he said. “Your daughter could live on the ranch next door, and that could still be an hour away.”

Moreover, the videoconferencing solutions behind telestroke — and much of post-COVID work and socializing — are not yet adapted to the heights Usmanov traveled. For about two-thirds of his ride, his phone had no connection at all, and simple sites could take minutes to load even when service was available.

Again — even with the benefit of sick leave, broadband internet, modern roads, and denser neighborhoods — Americans still tend to begin treatment too long after a stroke’s onset. So Usmanov has some justification for arguing that we need to focus on the first phase of stroke’s unforgiving timeline: between onset and calling 911.

For years, the public health establishment has tried education paradigms. Since 2013, they’ve used acronyms like F.A.S.T. to help laypeople recognize the telltale signs of stroke. But further study found that 30 days after an educational video, many participants had forgotten the symptoms, and that rates of timely 911 calling declined as well.

And in the Rockies, as in many lower-income precincts, there’s an element of class at play. “Generally, I met people who have done highly laborious work for years: construction, ranching, in the service industry,” Usmanov said. “And they tended to tell me similar things: ‘I’ve been in pain my whole life.’ ‘I broke something, but I had to work through it.’ Or ‘I experience weird sensations as part of my job: tingling, dizziness, numbness.’ Those are just the demands of the job — but they’re also symptoms of stroke.”

In January, Usmanov was the lead author of an American Heart Association paper finding that patients who were with family or trusted bystanders at onset tended to arrive at hospitals earlier than those who weren’t — or even than those who were with large crowds of strangers.

And his Great Divide ride underscored how much community matters in unforgiving terrain. 

Some nights he’d camp serendipitously with other racers, for camaraderie as much as for safety. He made a “good friend” in one fellow racer, only to watch him drop out after developing nerve damage: the consequence of biking over bumpy dirt track with inadequate suspension.

And 500 miles from the finish line, Usmanov was contemplating dropping out after a crash and without water. “A rider named Bob passed me, gave me a little Gatorade, a kind of a pep talk, and went on his way.” (Only about half of the roughly 260 racers completed, with Usmanov middle of the pack.)

Usmanov’s ride had one last goal in mind: it served as a fundraiser for the Global Brain Care Coalition, a nonprofit that shares his ambition to find preventative care for neurological ailments. (It has raised $20,000 at time of publication.) And the Rockies taught him that there may be promising interventions to be made between individual vigilance and institutional change.

“Spouses, neighbors, church members can take blood pressure together, can discuss health behaviors — and that will do a lot more than this ad hoc approach, of one individual at a time,” Usmanov said. “We’re social creatures, and that’s where we have to target change.”

It’s a fitting takeaway for a sociologist after a long, lonely ride through the mountains.

The question is no longer whether young people are using AI for mental health care, but the soundness of the advice they receive, health officials say.

A recent survey showed that the use of AI chatbots for mental health care among adolescents and young adults has increased 60 percent in just a year, a trend that survey organizers said is likely to continue.

Hao Yu, Harvard Medical School associate professor of population medicine at the Harvard Pilgrim Healthcare Institute, led the study, which had 1,727 participants.

Yu said as AI helps fill a treatment gap left by a rising problem among young people in the U.S. amid a shortage of mental health professionals, the focus should now turn to ensuring the counsel being offered is helpful and not harmful. The interview has been edited for clarity and length.

---

Gazette: What did your survey find?

Hao Yu: We did a similar study last year, and at that time we found one in eight young people, between the ages of 12 and 21, use AI for mental health problems. This year we found an even higher proportion, almost one in five. That means nearly 20 percent of young people in this country are using AI for mental health advice. That’s a rapid increase in one year.

Not surprisingly, we found variation in AI use across youth groups, with relatively older youth more likely to be using AI, and females more likely to use AI for mental health issues.

It may surprise some that we found the majority of youth using AI did not disclose their use to any adults, either parents or physicians. That might be concerning, because there is misinformation available through different AI models, which are not designed to provide mental health advice.

So, we would encourage parents and physicians to talk with their youth regarding their potential use of AI. They may be able to identify if their child has a mental health issue or help their child distinguish between accurate information and the misinformation we know is available online.

What does this tell us about the state of young adult and adolescent mental health care today?

This definitely shows that there’s a big issue facing the U.S. today. On one side, we have a rapid adoption of AI across different parts of our society. On the other, there is an increased prevalence of mental health issues among young people. Putting these two together makes it not so surprising when we see that youth are increasingly turning to AI for mental health advice.

Does it also highlight that there’s not enough capacity in our mental health care system?

Our healthcare system has an acute shortage of psychiatrists, especially child psychiatrists, and only half of those with mental health needs actually receive mental health treatment.

So, if a youth has mental health issues, AI offers some advantages. AI is patient. AI is available anytime, 24/7. You don’t need an appointment to talk with AI, and AI is nice. It’s rare for AI to be harsh or become impatient. So there are advantages, especially for people living in rural or remote areas with severe mental health workforce shortages.

How much do we know about the quality of the responses that come from these general-use AI chatbots? Is it still a concern that that kids and young adults may get bad, even harmful advice?

That’s still a concern. And it’s true that tragedies like suicide attempts have happened when youth talked with AI about mental health issues. AI companies promised they would install guardrails but there’s a lack of empirical evidence about how those promises are being kept.

Beyond this specific study, many researchers are interested in assessing the safety and efficacy of mental health advice offered by AI. More people are paying attention, so we’d hope there would be some improvement, but that remains to be seen.

The paper mentions that these AI models are designed to be friendly, agreeable, and flattering. Might kids also be getting information that is reinforcing in a positive, but still unhelpful way?

Most AI models try to be appealing, flattering, as a way to keep their customers, and many people, especially youth, easily hooked. But it’s still a concern that the advice, the information provided by AI, be really useful.

In our study, we found that the vast majority of youth, over 90 percent, say the advice is useful, but that’s still an empirical question we need to assess: Is AI-provided mental health advice really helpful and accurate?

The safety and efficacy of AI models is a priority for research because, as we said in the beginning, these things are already happening, and the trend among youth is increasing. As a society, we have the responsibility to make sure that what they are receiving is accurate and helpful.

You mentioned that there are AI chatbots that are designed to be therapeutic. I’ve seen earlier studies that show that they can be very good, but their development has been slow. Is a key message that we need to accelerate the development of these therapeutically trained bots?

Yes. We have an acute shortage of child psychiatrists who can meet unmet needs among our youth, so we want AI models that can mitigate the shortage. It’s not surprising, though, to see people concerned about the potential quality and safety implications of even the specifically trained, mental-health-oriented AI.

That’s why some policymakers, as in Illinois, have taken actions. Ultimately this is an issue of patient safety.

Where do you see this trend going? We’ve seen an increase from one in eight adolescents and young adults using AI for mental health needs to one in five. Are you planning on doing the survey next year, and do you expect it to be an even greater percentage?

We definitely will have follow-up surveys, and most likely will see an increased number. It’s not surprising given that AI is now part of daily life for youth, and that there are unmet mental health needs. AI is readily available and endlessly patient.

Were there any surprises in this for you?

One surprise is that we didn’t find a significant difference between rural and metropolitan areas, even though we know that a rural-urban digital divide exists and that rural and remote areas have a severe shortage of child psychiatrists.

If confirmed by future research that rural and urban kids are equally likely to use AI, that might help reduce the adverse effects of a workforce shortage in rural areas. AI has that potential for sure. But we want to make sure these specific AI models meet clinical standards for mental health advice.

The research described in this interview received funding from the National Institute of Mental Health.

Seventy-nine percent of Gen Z want a job mainly focused on helping people or making a positive difference in others’ lives. But only about half of workers in that 18-29 age bracket who want such a position say they have one. 

Those are the findings of a new report from the Harvard Graduate School of Education’s Making Caring Common project, in collaboration with Gallup and the Walton Family Foundation. The survey examines what members of Gen Z want from their work and how it affects their emotional well-being.

The data, taken from a representative sample of 2,436 people born 1997-2012, goes against popular stereotypes, according to Richard Weissbourd, faculty director of Making Caring Common and a senior lecturer at HGSE. 

“Gen Z is often characterized — in my mind, cartooned — as selfish and lazy, as not wanting to really invest in anything deeply. So I was really surprised, partly because of this misconception, to see that 79 percent of our respondents report that they wanted to have a job in the future that was mainly focused on helping other people,” Weissbourd said. “It was just a very high percentage. And it’s not just talk. Many Gen Zers are taking concrete actions based on their desire to do good, according to Deloitte surveys, such as choosing companies based on their social impact.”

Weissbourd explains why helping Gen Z find meaningful work could have a big impact both on the culture and in the lives of the young adults — and why there may be some good news on the horizon. The interview has been edited for clarity and length.

---

Gazette: What kinds of jobs do Gen Z find meaningful? 

Richard Weissbourd: We asked two different questions: whether they found their jobs meaningful, and whether their job was primarily focused on helping other people. 

Among Gen Zers who told us they found their jobs meaningful, about twice as many (84 percent) worked in jobs focused on helping people — such as jobs in healthcare, education, and the military, as those who did not work in these kinds of jobs (44 percent).

In response to an open-ended question about where they find meaning at work, we also found that many Gen Zers focused on those aspects of their jobs that helped others, such as a dishwasher who finds meaning in playing an important role in providing people food or a retail sales worker who finds meaning in helping people find clothes that make them feel good about themselves.

So I think two things are true. One is that a lot of young people want a job that’s directly helping others, like a job in healthcare or social work or teaching. But also, many young people are finding meaning in the aspects of their jobs that involve helping others, even if it might not be as obvious. 

This research only spoke to Gen Zers. So this can’t really tell us how Gen Z compares to other generations, is that right? 

That’s correct. From this and other data, there isn’t clear evidence that this generation is more altruistic or more socially active than other generations in general, but there certainly isn’t clear data showing that they’re less altruistic or less socially active. 

What’s different among Gen Zers who report feeling purpose and meaning in their lives? 

Over half of Gen Zers report mild (30 percent), moderate (18 percent), or severe (10 percent) anxiety and depression, and a whopping 70 percent of Gen Zers over 18 report mild to severe anxiety and depression. But those who say they have felt a sense of meaning or purpose in their lives in the past month are much less likely to report symptoms of anxiety and depression than those who do not feel their life has felt meaningful or purposeful. 

Meaning and purpose are the deep engines of our lives. They provide this deep gratification that’s beyond hedonism, beyond temporary joy. We can talk about banning cellphones in schools, which I think is a good thing, and providing more therapists, which I also think is a good thing, but my guess is that helping other people and making a positive difference in others’ lives is significantly more likely to strengthen Gen Zers’ mental health than those other strategies.

You found that just over half of Gen Z adults (56 percent) who want a job mainly focused on helping people have one. When asked about perceived barriers to taking those jobs, many of them said they perceived care-oriented jobs to be low-paying and emotionally draining. Are they right? 

They are right in the sense that a lot of these jobs aren’t affordable. I mean, if you want to be a childcare worker or preschool provider in a city like Boston or New York, it’s very tough. We need to figure out ways to pay people more, reduce housing costs, provide loan forgiveness for people in helping professions in under-resourced areas, et cetera.

Certainly there are helping jobs in ed tech, philanthropy, and other industries that pay well. But we should also be creating-less expensive pathways into careers in healthcare and education. 

The emotionally draining aspect is interesting to me. These jobs can certainly be emotionally draining, yet in our data, when people feel they’re making a positive difference in other peoples’ lives, they also report less anxiety and depression. Teaching, for example, can be very depleting, but it’s also very gratifying for a lot of people. Both of these are true. 

So is it fair to say that Gen Zers see helping jobs as a nice-to-have in principle, but only as long as they can pay their bills? 

It’s possible. Here’s one finding that speaks to that question. We asked people: “You have a job that provides you with a feeling of meaning and purpose. You’re offered a new job that has double the salary but will have significantly less meaning and purpose. Would you take it?” 

To summarize, people were roughly split: 46 percent said they would take the higher-paying job, 28 said they would keep the meaningful job, and 26 said it would depend on the specifics. 

Now, when we dug a little deeper, we found that perceived financial security clearly mattered. No matter their actual income, adult Gen Zers who felt financially insecure were more likely to take the higher-paying job (53 percent) compared to those who were living very well (41 percent). But a significant number of people with low financial security are also taking up helping jobs. So it’s a complicated picture.

With an aging population in need of care and potential AI disruptions to knowledge-economy jobs, is this a rare moment where this population’s desire for helping jobs could align with where the economy is headed? 

Yes. I’m not an expert in the economy, but I can say a very common view is that we’re shifting from an economy that values information and knowledge to an economy that values interpersonal skills, social-emotional skills, and the capacity to care for other people.

In healthcare, for example, you can imagine AI taking over a lot of the knowledge functions that a hospital might rely on, freeing up more people to do really high-quality patient care. The gist of it is that this is a time when what young people are interested in and what the economy demands may rhyme.

Economists and Wall Street analysts frequently warn about what sounds like an immutable law of economic gravity: Growth that gets too hot for too long risks provoking an inescapable cool-down known as a recession.

But that notion is wrong, according to a new book about the history of recessions in the U.S. and U.K. dating back to 1700, “Recession: The Real Reason Economies Shrink and What to Do About It.”

Contrary to popular thinking, these “busts” are not the inevitable result of flaws in the economic “booms” that immediately precede them, and they’re not cyclical, or necessary, course correctives. In fact, recessions have relatively little effect on the economy over time compared to periods of growth, says the book’s author, Tyler Goodspeed ’08, M.A. ’11, Ph.D. ’14.

Economic historian Goodspeed discusses what triggers recessions and why war is the most prolific “serial killer” of economic expansions over the last 400 years. He served as acting chair of the White House Council of Economic Advisers during President Trump’s first term and is now chief economist at ExxonMobil. The interview has been edited for clarity and length.

---

Gazette: What’s the definition of a recession?

Tyler Goodspeed: I adhere to the definition used by the National Bureau of Economic Research, which is a broad-based economic contraction that’s spread across sectors and that lasts for more than a few months.

I think armed only with a graph of the unemployment rate, one could do a pretty good job picking out historical recessions, which are typically characterized by the substitution of a sudden, sharp upward movement in the unemployment rate for a more gradual downward or sideways movement.

You say the conventional wisdom around why recessions occur is wrong. How so?

The stories we tend to tell ourselves about recessions is that there was something fundamentally wrong in the expansions that preceded them. There was some error, some excess, some imbalance, to which recession was the inevitable, maybe even necessary, outcome.

Fortunately, that view yields several empirically testable and falsifiable hypotheses.

First of all, if there was something wrong in a so-called boom to which the bust was the inevitable remedy, then we should expect expansions to die of old age. So the longer an economic expansion has been going on, the more those errors and excesses and imbalances accumulate, the likelier it is to die.

We would expect that higher, faster, longer economic expansions should be followed by deeper, faster, longer recessions.

We should expect that the frequency of recessions should be roughly constant across time and space.

We should expect that the economy should look fundamentally different at the end of an economic recession and subsequent recovery from how it would have looked in the absence of the recovery of the recession.

We should also expect that the increased role of the state should have attenuated the depth and duration of recessions, and we should also expect that recessions should be in some sense predictable, that there should be some indicators in the height, the speed, excess, whatever, in an economic expansion that should predict recession.

In reality, when I analyzed these 132 recessions going back to 1700, all of those hypotheses are rejected by the data. All of them. Expansions don’t die of old age; deeper, faster, longer recessions don’t follow higher, faster, longer expansions; recessions have become less frequent over time.

The U.K. has long been less recession-prone than the U.S.

Typically, economies look pretty similar at the end of a recovery to how they would have looked in the absence of a recession; and recession depth and duration have been remarkably constant over time.

And finally, they’re not predictable.

So, all those hypotheses generated by the boom/bust view are just rejected by the data.

Do recessions share any common patterns historically?

There are some things that are roughly constant across time and across space. The depth and duration of recessions have been pretty constant over time. Most recessions are over in about a year; the vast majority in fewer than two.

A key feature of recessions is a sudden sharp upward movement in the unemployment rate, and that’s typically driven in the first instance not by a sudden increase in the rate at which firms lay off workers, but rather a sudden collapse in the rate at which they hire workers.

Are they in any way avoidable or inevitable?

Recessions will continue to happen because history will continue to happen. There’s never been an immortal economic expansion.

One of the optimistic findings in the book is that recessions have become less frequent over time in both the U.S. and the U.K. That is a long-run structural trend going all the way back to 1700 with no clear break at any specific points in time toward longer-lived economic expansions.

What’s driving that is households, businesses, and consumers have been learning how to better absorb the kinds of shocks that historically would have generated recessions. Two examples: Our banking systems are generally more resilient, more diversified today than they were in the 19th century, and our energy systems have become a lot more diversified, both across fuel types and within fuel types.

You argue there are three categories of “shocks” that typically precede recessions. You call them “Acts of God,” like pandemics or severe weather that slows commerce or damages crops; “Acts of Man,” such as major bank fraud; and “Acts of Church,” which are state actions like imposing consumer credit controls. But you say there’s one thing that’s been “a prolific serial killer of economic expansions” over the last 400 years: war. Why is that and is that true for all armed conflicts?

Not all wars are created equal. What I’m talking about is either world war or war that visits your soil in a profound way.

Take some of the longest recessions in history for the U.K. and the U.S. The U.K. had a very protracted recession that began in 1943 during World War II and continued all the way till 1947. They also had a very long recession at the end of and in the immediate aftermath of World War I. Another very long recession during the Seven Years War in the 18th century.

In the U.S., we had a Depression-magnitude recession during the American Revolution, and in the immediate aftermath of the American Revolution during a very volatile period.

One of the most fascinating recessions I found looking back to 1700 was in the North American colonies from 1717 to 1720. This was the peak of the so-called Golden Age of Atlantic piracy.

You had the end of a major European war during which the British government had paid private sailors to become privateers. So, at the end of the war, there were all these unemployed sailors, and they turn to piracy.

Trade and commerce in the American colonies just ground to a complete halt. It was only after some of the most infamous pirate names were killed or captured or accepted the king’s pardon that you had a resumption of normal commercial flows.

Why is there always so much fretting currently over the possibility of a recession, given that they tend to be relatively brief and less frequent? It’s especially puzzling since the U.S. notched two of its longest economic expansions ever in the last few decades, and expansions exert greater overall influence on an economy.

There’s a lot of fretting because our minds don’t deal well with randomness, and at the end of the day recessionary shocks are random. So, we’re constantly searching for predictive patterns to recessions and embed those perceived patterns in parables of boom and bust.

But the moral, boom-bust stories we tell ourselves about recessions are not entirely innocuous because if it’s the case, as I find in the book, that economic expansions die healthy and innocent, then there is a risk that we as policymakers are perpetually seeking to sedate or otherwise overmedicate economic expansions that fundamentally die healthy and innocent.

And typically, a few years on from a recession, economies look pretty darn similar to how they would have looked had they continued uninterrupted along a long-run trend.

You touch upon a really important observation, which is that the majority of years in which economies expand matters more for long-run prosperity and indeed human flourishing. And so, we probably ought to be at least as concerned about raising that rate of trend growth as we are about avoiding recessions, because ultimately the periods of expansion matter more.

Despite recent “shocks” to the U.S. economy like tariffs, the war with Iran, the spike in fuel and energy prices, and workforce reductions attributed to AI, the country is not in a recession. What signs from history do you look for to help forecast whether one is on the horizon and how detrimental it might be?

Something that I learned is that it takes either a really big shock like a massive pandemic or a cluster of smaller shocks to push a large, diversified economy into a period of outright contraction. It takes a lot.

In terms of what I look for, recessions are typically characterized by a sudden sharp increase in the civilian unemployment rate. That’s the best real-time indicator of whether you’re in a recession. It can’t predict whether you’re going to be in a recession, but it’s a good indicator of whether you are.

I find that recessions are unforecastable. They’re unpredictable. However, what the book does equip one to do is to get a sense of the kinds of shocks that historically have contributed to recession, and then the reader can ask themselves, do I think the conditional probability is higher or lower than the unconditional probability? And in the past century, there’s been, in any given year, about a 1-in-100 chance of a pandemic-related recession; there’s been about a 1-in-10 chance in any given year of an energy-related recession in the U.S.

When exploring different sides of an issue, a devil’s advocate can play-act an opposing perspective, but the exercise has limits. It’s better to talk to someone who genuinely disagrees with you.

“You actually have to make the effort to get people in the room who deeply believe in, think about, and live by a certain set of aspirations or ideals,” said Jenna Silber Storey, co-director of the Center for the Future of the American University at the American Enterprise Institute, a conservative think tank. When faculty members discuss intellectual diversity on college campuses, it’s not enough, she said, to “theoretically entertain what political conservatives might think.”

Dean Tomiko Brown-Nagin of the Harvard Radcliffe Institute agrees. When she convened a workshop on intellectual diversity last November, she and her co-leads, Harvard Ed School historian Julie Reuben and Keith Whittington of Yale Law School, took care to choose scholars and leaders whose wide range of political and ideological perspectives would put their commitment to collaboration to the test.

The group proved up to the challenge, in May publishing a white paper Brown-Nagin described as “a true negotiated document that reflected consensus.”

Storey, one of the 15 authors of the paper, credited the success of the effort in part to Radcliffe’s insistence on a truly diverse mix of thinkers.

“Simply the fact that the Harvard Radcliffe Institute invited a significant number of political conservatives to discuss this question was astonishing to many people on the political right,” she said. “It just signals right off the bat: We’re going to take your opinion seriously; we’re inviting you to be equal partners in drafting this statement.”

In the paper, “Intellectual Diversity and the Mission of Higher Education,” the authors came to several shared conclusions. They agreed that universities are facing a legitimacy crisis, in part due to public skepticism of higher education’s civic value. The tendency of faculty and students to skew more liberal than the general population is one important contributor to that skepticism, the authors found.

More concerning, “students and faculty do not feel fully free to express themselves on certain issues,” the white paper states. The group advocated for programs on campus to replace a “censorious culture with one that is more open to debate and disagreement,” adding: “We are largely united in the view that most colleges and universities simply do need more ideas and methodologies often associated with political conservatism in order to achieve their mission.”

While the group discussed areas from research to governance, they agreed that they didn’t have as much information as they needed to both diagnose and solve problems. “In the absence of rigorous evidence, it is exceedingly difficult to define precisely the problem we face and to make progress toward solving it,” the document says.

As an example of the research that would help, the authors highlighted surveys and analyses performed by workshop participants Neil Gross, a sociologist at Colby College, and Jon A. Shields, a government professor at Claremont McKenna College, and called for more study of topics such as course content and teacher and student experience.

As an outgrowth of the white paper, Brown-Nagin said that Radcliffe will soon announce a research project that will include surveys of faculty and students to clarify “the extent of political skewing on university faculties, why that might happen, and most important what, if anything, it means for classroom dynamics.”

The group also used the paper to highlight where they could not come to an agreement. Though they reached consensus on wanting a wider range of perspectives, they disagreed about whether universities should hire directly for ideological diversity.

Brown-Nagin, for one, questioned the idea that a faculty member’s political orientation inherently deprives students of a full range of ideas. Part of what people have attacked as politically skewed teaching, she said, is likely ineffective teaching — “faculty not having the tools to actually facilitate a conversation about a hot-button topic in class.”

“If what passes as political indoctrination is just a matter of faculty not having the right teaching tools, we can give them additional tools so that they can teach more effectively,” she said. At the Bok Center, she noted, Harvard offers courses on promoting healthy classroom dialogue, which are now mandatory for teaching fellows in the Faculty of Arts and Sciences.

Along the same lines, students’ reticence to speak on controversial issues might not be the result of faculty pressure. The 2024 report from Harvard’s Open Inquiry and Constructive Dialogue working group, co-led by Brown-Nagin, “found that fear of student judgment, rather than faculty retribution, is the primary motivation for most self-censorship,” she said.

“It’s important both to train faculty and to cultivate the right mindset and teach the right skills to students,” she added. She hopes that some of the new educational modules for incoming students across the University can help prepare them for deeper classroom discussions.

Reuben stressed that, while contemporary debates about academic freedom criticize universities as overly liberal, the ideological makeup of academia has shifted over time. For much of the 20th century, she said, “there was, if anything, less tolerance for what we might consider the progressive end of the spectrum than the conservative end of the spectrum.” It took decades, beginning in the 1960s, for a more progressive faculty to ascend.

Brown-Nagin expanded on the same point.

“Concerns about political indoctrination stretch far back into our history,” she said, from promoting certain narratives about slavery and the Civil War in the early 1900s to McCarthyism decades later. “There have long been fears that faculty are teaching the wrong subjects, that these politics are invading the classroom. And these concerns span the political spectrum.” It’s important, she said, to appreciate that the pendulum swings in both political directions — and that that pattern will likely continue over time.

As conversations about intellectual diversity continue, Reuben hopes that people pay attention to the paper’s emphasis on promoting professional norms among faculty. As a historian, she draws inspiration from the American Association of University Professors’ 1915 Declaration of Principles on Academic Freedom and Academic Tenure.

The white paper quotes the principles directly:

If this profession should prove itself unwilling to purge its ranks of the incompetent and the unworthy, or to prevent the freedom which it claims in the name of science from being used as a shelter for inefficiency, for superficiality, or for uncritical and intemperate partisanship, it is certain that the task will be performed by others — by others who lack certain essential qualifications for performing it, and whose action is sure to breed suspicions and recurrent controversies deeply injurious to the internal order and the public standing of universities.

In other words, Reuben said, faculty need to “step up and take responsibility for these things” — or others will do it for them.

Artificial intelligence is poised to soon change virtually every part of our lives, including the way we lead.

“There is no policy area, there is no function in the public or social sector, that isn’t going to be massively impacted by the accelerating power of generative AI,” said Jeremy Weinstein, dean of Harvard Kennedy School. “We need our curricular offerings to rise to meet that challenge.”

HKS recently unveiled a series of new initiatives across a range of areas, including financial aid and several AI and technology-focused shifts as part of a broad, 10-year strategic blueprint for the School.

As part of the plan, HKS will launch a new tech policy concentration for its master’s in public policy candidates and has hired two AI and policy professors, including MIT’s Stephen Casper, who joined HKS this summer. Casper will teach methods courses and electives on AI governance. The School also plans to look at offering a tech policy degree.

Some changes are already underway. This spring, students were offered two semester-long courses, one on the national security implications of AI led by Jake Sullivan, former national security adviser to President Joe Biden, and another on AI and policymaking that Weinstein co-taught with Sharad Goel, an applied mathematician and computer scientist at HKS.

An important focus involves getting the students more comfortable with the technology by learning how to use AI assistants to code and conduct rigorous analytics.

The goal is to ensure that they can use the powerful tools to solve public policy problems “and not just be observers or passive consumers” of AI, said Weinstein, Don K. Price Professor of Public Policy at HKS and a government professor at the College. “And that is just getting easier and easier because of the skill of the tools themselves.”

AI has “no clear analogue in history,” said Sullivan, now the Kissinger Professor of the Practice of Statecraft and World Order at HKS. Unlike other national security technologies, such as the internet or the nuclear bomb, AI development is not being led by government, but by the private sector. That makes this moment an especially “exciting” time of “radical uncertainty” for those entering both the public and private sector.

Weinstein and Goel’s class, “Public Problem Solving with Generative AI: Challenges and Opportunities,” drew inspiration from a multidisciplinary course that Weinstein, who came to Harvard in 2024, had taught at Stanford.

Combining case study and discussion with hands-on assignments, the course objective is to help HKS students, many of whom were not STEM concentrators in college, understand the technical underpinnings of AI while confronting the big policy questions this powerful new technology raises across domains from national security and science to education and healthcare, said Goel, who developed the School’s AI module two years ago.

So far, HKS students have responded enthusiastically.

Given AI’s importance, it’s essential to not only understand how this technology is being deployed globally but also how AI “works under the hood,” said Jacob Gordon, M.C./M.P.A. ’26, who took Weinstein and Goel’s course and valued the opportunity to vibe code and use Python as part of class assignments.

Gordon, who hopes to start an AI safety nonprofit, supports the push to expand and deepen the School’s focus on AI and technology.

“It’s vital, and HKS needs to move fast if it wants to contribute to this small decision-making window as the world figures out what path to take on the future of AI,” he said.

In one lab, students in Weinstein and Goel’s course tackled the challenge of getting 911 callers quickly routed to the appropriate agency or service provider when the situation doesn’t require law enforcement.

Some emergencies, like a behavioral health crisis, may demand other expertise, and dispatching armed police officers to a scene where they’re not needed may add danger to volatile situations, said Goel.

But preparing 911 call takers to quickly and accurately assess the type of help to dispatch often requires two or more years of training, a costly process that can strain municipal budgets, he added.

So, using generative AI, students built voice-training tools that used real-world data to simulate 911 callers.

The simulators tested call takers’ decision-making skills and later assessed their performance, representing a potential low-cost option that can better identify training deficits and get new call takers online faster.

“Few classes give you the chance to discuss policy issues with someone who was directly involved in shaping those policies at the highest levels of government just months earlier.”

Sarah Zheng, M.P.P. ’26

Globally, AI has already begun to remake major policy and geopolitical questions involving national security and AI regulation, said Sullivan, who served in the Obama administration as an adviser to then-Vice President Joe Biden and deputy chief of staff to Secretary of State Hillary Clinton.

“The developments of this technology are shaping relations among nations; they are shaping the national development strategies and growth strategies of nations; they’re shaping the power balances among nations” and affecting “every aspect of national security, economic security, and social interaction,” he said.

His class focuses on how nations, particularly the U.S. and China, are trying to take advantage of AI’s potential while managing its risks.

Sullivan said during his time in Washington, few who were not working on the technology side of government focused on how AI might reshape their policy domain, let alone pursued training in how to use it, but that’s no longer the case.

“You don’t need everyone to be a technical expert, but I do need everybody to have a framework and a sense of priority for this because it’s going to be so central to the big questions,” he said.

Sarah Zheng, M.P.P. ’26, who reported on Chinese technology and geopolitics for Bloomberg News and the South China Morning Post, said Sullivan’s class offered firsthand insight into how major decisions on AI and national security are being made and presented a full spectrum of perspectives on AI, from skepticism to optimism.

“Few classes give you the chance to discuss policy issues, such as technology export controls, with someone who was directly involved in shaping those policies at the highest levels of government just months earlier,” she said.

Weinstein said when HKS first began to incorporate technology into the M.P.P. curriculum several years ago, the intent was to prompt students to think about “everything from how is technology used in the defense sector; how do we think about cyber issues and cyber threats; how do we think about civic technology … and how do we govern in a digital age.”

Now, the School’s growing emphasis on AI and technology reflects the inevitable reality as well as the potential promise and peril of what they may bring.

“For me, it’s very simple. One of the most important things that we can do as a school over the next generation is to populate public institutions, including people in political leadership, with the technical understanding and know-how to take advantage of this technological revolution to improve the ability of our public institutions to deliver what people need,” Weinstein said. “And this kind of teaching is essential for that.”

Sarah Newman is collecting secrets. Specifically, she’s collecting secrets about AI. 

At conferences and other events, she puts out a wooden box and invites passersby to anonymously disclose the ways they’re using the powerful, mushrooming technology.

“I’ve written a speach [sic] with AI on a topic I don’t have a clue about,” wrote one anonymous confessor. 

“Using AI to write official policy language without necessarily disclosing,” wrote another. 

Newman, director of art and education at metaLAB at Harvard, sees a lot of uncertainty and angst out there. The aim of her project is to promote discussion about the personal use of AI at a time when its spread has already outstripped the ability of businesses and institutions to define what is appropriate in every instance.

“People don’t know what the rules are,” she said. “There are not established shared norms around both AI use — what’s appropriate and what’s not, across contexts — including when you should disclose use.” 

Since the 2022 release of the popular Large Language Model ChatGPT, businesses, schools, and universities as well as newspapers and publishing houses have been trying to sort it out.

But those efforts have been complicated by the technology’s defining property — its ability to mimic human output. This means it can be difficult, if not impossible, for institutions to police whether users are cheating. 

And there have been recent high-profile accusations of misuse.

In May, the British literary magazine Granta responded to allegations that one of its prize-winning short stories, Jamir Nazir’s “The Serpent in the Grove,” was AI-generated. (Nazir has denied using AI to write the piece.)

Separately, the author of a book about truth in the age of AI acknowledged using the technology in writing his book after The New York Times identified misattributed or fabricated quotes in the text.

In addition, news outlets and social media have produced a steady drip of stories about the use of AI in wedding vows, dating app messages, or eulogies.

The eventual societal fallout of all this uncertainty is unclear. But fields like journalism already have been affected.

“What we’re actually seeing is audiences assuming that real media is AI-generated, and that’s a really concerning trend to me,” said Andrew Deck, a reporter who covers AI in the industry for Harvard-affiliated Nieman Journalism Lab. “It indicates that [for] some audiences, in some ways, their default assumption these days is that something is false rather than the other way around.” 

In truth, Deck said, very little content at reputable news outlets is generated using AI. Reporters mainly use AI for things like transcribing meetings, scrounging up sources, or iterating title ideas. But they’re doing so amid historically low trust in news organizations.

Pew reported last year that about 56 percent of American adults had “a lot of” or “some” trust in information from national news outlets, down 20 percentage points from 2016. AI, with its reputation for spouting hallucinations with certitude, presents a challenge for an industry grounded in accuracy and pushing back against what Deck called a “trust crisis.” 

AI use labels, Deck said, “feel like a way to indicate to readers or audiences that maybe they should have an additional layer of skepticism.”

But disclosure might not always be a clean solution. “AI-assisted” could mean, in theory, anything from a quick source search on Perplexity to an editorial pass that was more like a rewrite.

Issues over uncertain standards appear to have even filtered down to families.

Leah Plunkett, Meyer Research Lecturer on Law at Harvard Law School and faculty associate at the Berkman Klein Center for Internet and Society, is an expert in digital technology, youth, and families.

In her work with K-12 schools, she’s noticed that kids are on high alert for signs that adults in their lives have used AI when they themselves have been forbidden it. 

“They have an antenna, as kids and teens — particularly teens — for potential hypocrisy or double standards,” she said. “It can be hard to get them to understand that actually different roles in a school or community ecosystem have always been permitted, and at times required, different tools and resources, because they do different activities.”

Newman of metaLAB has been fascinated by secrets, and why we feel the need to keep them, for a long time. A 2016-2021 project examining why we share so much with our digital devices, “The Future of Secrets,” was exhibited internationally. She moderated a metaLAB panel on AI, secrecy, and disclosure this spring. 

Newman believes that norms for AI use will become clearer in the coming months and years. She added, however, “We shouldn’t expect norms to just emerge. We should reflect on our own use and then help shape the norms we want to see.”

Like many people, Newman said she has a love-hate relationship with AI. But, she said, “People shouldn’t feel a need to lie. We should be curious about — even intrigued by — our own mixed feelings, and work toward normalizing appropriate, responsible AI use.”

Newman credits her former research assistant, Olivia Tai ’22, with the idea for the AI secrets project.

Tai is an artist and facilitator who has published her own collection of anonymous confessions. She hopes the project can encourage people to talk more openly about their use of AI and be more curious and understanding of how others view the issue.

“Like with all taboo topics, the more we shame each other, the less productive our conversations become,” she said. “We stop hiding when we feel some acceptance, understanding, and recognition from others.”

On May 29, the Office of Management and Budget (OMB), along with federal grantmaking agencies, proposed an overhaul to the government-wide framework that has governed for more than a decade how federal grants are reviewed, awarded, and managed.

The proposed rules to the Uniform Guidance, the framework’s formal name, would significantly change the way federally funded research is selected and overseen, giving political appointees a larger say in funding decisions and expanding the government’s ability to cancel grants already underway.

The proposal drew nearly 500,000 public comments before the window closed on July 13, with a large majority opposing the new rules. A final rule could come as early as Oct. 1.

Harvard submitted comment letters from the University, the Medical School, and the T.H. Chan School of Public Health, which urge OMB not to move forward with the proposed changes.

The Gazette spoke with John H. Shaw, senior vice provost for research, who wrote the University’s comment letter, about what the proposed changes would mean for researchers and for the federal research partnership. This interview has been edited for clarity and length.

---

What has historically been the purpose and scope of the Uniform Guidance? Can you give us an overview as to what’s happened recently?

The Uniform Guidance that’s issued by OMB is a framework that governs how the federal government allocates taxpayer dollars to fund its grant-making to research universities like Harvard, to other nonprofits, and to a wide range of organizations across the nation that the federal government seeks to support.

This framework represents the “rules of the road”: It ensures that those dollars are spent efficiently and effectively, and are consistent with the federal government’s expectations in providing that resource.

The scope and the breadth of the proposed changes, and the broad community they would impact has, I think, inspired such a strong response.

It’s typical when a proposed change to the Uniform Guidance is issued that the federal government calls for comment so that the various stakeholders have an opportunity to respond. The government, in turn, has a legal obligation to respond to those major criticisms — either to revise the policies or proposals or to explain why they’re not going to do so.

How do the proposed changes alter the current framework? What is the scope of these revisions?

They are vast in scope. They would affect more than a trillion dollars of federal spending across the range of areas in which the federal government awards grants and resources to help support the general public.

The document is hundreds of pages long. It’s very dense, and as a result, it reflects an amalgamation of a lot of different things that the administration has sought to accomplish through executive orders and other means over the last two years. 

There are areas that represent a fundamental change in terms of how research is conducted at universities, at affiliated hospitals, and other types of entities. That is ultimately what has motivated Harvard to respond in a way that is thoughtful and responsible. There are a couple of areas that really are the underpinnings of the ecosystem of innovation that the federal government has supported over the last 75 years in higher education. 

That innovation ecosystem is the reason that U.S. universities are world leaders in so many areas of technology and science more broadly. People feel as though these changes, if implemented, threaten to significantly reduce the competitiveness of the United States. They could have very direct and immediate impacts on areas of national importance — like national security, the health and well-being of patients, and the development of crucial technologies that we need to maintain our competitiveness and continue to support a vibrant economy.

How has Harvard responded to these proposed new rules?

Well, naturally, Harvard would have responded to something of this scale and this potential impact, but that was actually strengthened by a chorus of interest and concern that came from almost every part of the University community — from our faculty, researchers, students, and  alumni — recognizing that these changes would have dire consequences for the U.S. research enterprise.

We were certainly motivated to plan a formal comment, and we did that by reaching out to the people in the University ecosystem who would be impacted. We received a tremendous amount of very constructive and thoughtful input, which we used to inform the University’s formal comment to these proposed changes. 

In addition, the Medical School and the T.H. Chan School of Public Health have submitted their own comments focused on the specific concerns of their communities. Many other individuals, as researchers, as faculty, have also submitted responses. And finally, we’ve aligned with major national associations that support higher education and research, like the Association of American Universities and Council on Governmental Relations, that also recognize the threats posed by these recommended changes and want to have a thoughtful and aligned response.

You mentioned two key focuses in the University’s comment letter, the first being the introduction of new political oversight of federal grants. Can you tell me more about this?

This is probably the most impactful change. The relationship that the federal government has had with higher education and other institutions to conduct research has been successful because it placed the merit of proposed work as the standard by which decisions are made to allocate resources. 

It’s important to clarify that the federal government and the various agencies that help support this science always determine what the priorities for the nation are. There are program managers and others who work in these agencies who consult with stakeholders and determine that these are areas that warrant federal investment, and those areas are not static. They change over time.

In my 30 years as a federally supported researcher, in my own lab, the priorities for national research have varied substantially, and that’s good. That’s healthy. That’s something that needs to continue to evolve. 

But what’s being threatened here is that once those priorities are established by a funding agency — that we want to prioritize research on a particular form of cancer, or on a particular area of technology — the decisions about what awards to make would be political, not based on merit.

That will have dire consequences. Making something political means that those priorities are going to change depending on the nation’s political leadership. Research is not something that survives well when it has to be turned off and turned on with great frequency. It needs to be sustained. 

I’d also say that merit-based peer review is a process that is very challenging to get through.

Many federal agencies that award grants have a less than 10 or 15 percent success rate. So, really, to get an award approved based on merit is a very high bar. It means that your proposal has gone through scrutiny, often in a highly specialized area of research, by the world’s leading experts in that field — it is a very strong process. 

The answer to any individual investigator isn’t always favorable. Most of the time it’s not. But it motivates you to work toward an even higher standard. This is the best way to allocate taxpayer money to achieve the goals of the agencies and the federal government.

The other focus you mention is federal agencies’ expanded authority to suspend or terminate grants. Can you talk more about this?

Research doesn’t often hinge on a single eureka moment. It’s usually something that needs to be sustained over years to make important breakthroughs and to innovate. The ability, once awarded, to have confidence that you can conduct that research over the term of the award — whether that be three years or five years — is critical, because you need to bring on staff and students to help enable that research, and we, as an academic institution, make multiyear commitments to our researchers. 

And so, if there is this prospect that political changes or whims will turn research on and off midstream, it limits the University’s ability to invest in, and sustain, the most important part of the research enterprise, namely the students, staff, and faculty who enable it.  

Frankly, this uncertainty will make faculty risk-averse. They won’t be able to bring in as many students, bring in as many researchers to conduct as wide a range of research, because they’ll have this threat hanging over them that their grants can be terminated. 

This would not include grants that are terminated for cause. That capacity already exists and should be maintained. The OMB’s suggested shift would permit midstream terminations that really reflect changing political winds, which would truly threaten the opportunity to fulfill the potential of the research that’s been supported.

This is not only a waste of taxpayer dollars, but poses the risk of long-term harm to research outcomes. If you fund something and only get 25 percent along the way before it’s canceled, we never have the potential to realize the benefit of that investment — and affected patients and the public never get to benefit from that research.

What are indirect costs? And what does the proposed rule say about their role in funding?

To many, indirect costs are one of the least understood aspects of supporting a research enterprise. Essentially, these are funds within each federal research grant that support the costs of the research’s associated administrative and facilities expenses. We have seen, over the last few years, threats to reduce indirect cost recovery to very small percentages.

What’s proposed in the OMB guidance is something different because, essentially, those attempts to just reduce it to smaller levels have already been stopped by the courts and by Congress. The OMB guidance is suggesting that a university’s indirect cost recovery rate is something that can tip the scales of whether that award should be made to that institution or another. 

In contrast to direct costs, like salaries, equipment, materials, and supplies, which are assignable to research activities on a specific project, indirect costs are other essential expenses that reflect the true cost of that work — costs like building infrastructure, laboratory facilities, electricity and utilities, and administrative staff to support administration and compliance.

Indirect costs are critically important to the conduct of research, but they aren’t easily attributed to a single project, because they support state-of-the-art facilities and labs and the infrastructure necessary to ensure that research is conducted safely and in compliance with federal and state laws and institutional policies, ranging from environmental health and safety to research security.

The facilities are where the work is done. Without them, work could obviously not continue. And of course, the administrative costs are necessary because we need to conduct our research in ways that comply with the rules that the federal government has set for us. And we’re committed to adhering to those because they are critically important — like how we treat human subjects who are engaged in research. 

The critical challenge with using this rate as a thumb on the scale in deciding to award funding is that the administrative part of this is already capped so the differences between institutions are really in the facilities.

If you start awarding federal grants preferentially to institutions that have lower indirect costs, it means one of two things: It means you’re awarding them to places where the cost of buildings and infrastructure is less, or it means you’re awarding them to institutions that are less willing to invest in their own research enterprise. 

At Harvard, a lot of the research we conduct is lab-based. It requires major University infrastructure investments to have the capacity to do the research, and indirect cost recovery is a way that we partially recover some of the costs associated with making those investments. And so, if we disincentivize making those investments, universities will make fewer of them, and will have less capacity to do research, and certainly less capacity to innovate.

In addition, we want to ensure that decisions to fund proposals are based on scientific merit and transformative potential, not what costs the least.

What problems could the new rules present for scientists and researchers?

Well, for the scientists and researchers themselves, it comes back to this issue of sustaining their work — whether they are running a lab and looking to hire research staff, or whether they are students who are trying to meet academic degree requirements and have the capacity to go on and start their own labs in other universities or other research institutions across the country. It’s a commitment, of course, that we make to those students when we admit them — that we’re going to be there to support them through that process.

Moreover, the subjects of this research are important, deepening our knowledge of the world, while also creating new technologies, medicines, systems of understanding and other insights that lead to practical benefits for people in their everyday lives.

If grants can be terminated based on changing political considerations, that creates tremendous fiscal uncertainty, and so we may be forced to admit fewer students. We may hire fewer postdoctoral fellows.

The lack of certainty regarding the commitment that we make to these researchers will also undermine our ability to recruit the best talent from around the world, which is so critically important to the U.S. research enterprise.

If you look at any university, including Harvard, the impact of international students and researchers is profound. It’s to the real benefit of our nation that we maintain the opportunity to attract the best and brightest, because they help enable this innovation, which obviously benefits our nation and our allies across the world. The kinds of changes proposed by OMB would sow unpredictability, and therefore they would clearly reduce the scale of our research enterprise, and therefore its impact.

Are there examples you can share as to how these proposed changes could slow or halt current research, and how that translates into impact? How do rules that disproportionately affect universities like Harvard harm our national scientific competitiveness?

For the second question, research is a collective enterprise. We at Harvard are the beneficiaries of awards from many federal agencies, and we take very seriously the responsibility to conduct that research thoughtfully and capably.

We also, through many areas of research, interact with other research universities and institutions across the country, and in some cases around the world. And again, that’s truly important to enable the best and brightest minds to come together to solve some of the most important challenges that we have in terms of breaking through barriers in biomedical research and advancing critical technologies like quantum science. 

The ability to arbitrarily terminate an award in one place will not be isolated in its impact. It will have a ripple effect through this ecosystem. Threats to universities like Harvard will be felt by our important research partners across the nation, which exist in almost every state, and of course span organizations beyond universities, like research hospitals. 

To address the first part of the question, there are just so many examples where the outcome of research that’s had a major positive societal impact wasn’t clearly foreseen at the inception of that research. So much of our fundamental research takes decades to come to fruition, and examples of that illustrate the need for a sustained commitment to the best and most promising fundamental research. 

Its potential can’t be foreseen by a political appointee or someone who is not intimately familiar with the area of science and technology. But someone with deep scientific and technical experience can — through a merit-based review process — assess whether this research will help gain fundamental understanding of a system, biological or physical, that allows us to decipher the code to unlock doors that are closed to us right now.

That will have value, not only in generally increasing our knowledge, but potentially enabling us to translate that research in impactful ways for society.

What are a few examples of this research?

A major revolution in biomedical treatments today results from our ability to do gene editing. Technologies like CRISPR, of course, are examples of one of the fundamental ways in which gene editing is enabled.

Some of the most exciting and impactful research in this area today had its foundations, at least in part, at Harvard University, through faculty at Boston Children’s Hospital who were affiliated with Harvard Medical School. In one example, our scientists helped develop a capacity to turn on and off fetal hemoglobin in ways that could allow us to treat the blood diseases sickle cell anemia and thalassemia. 

It took decades from the time that fundamental research was conducted to the point where these therapies were available to patients. Now, of course, their success has been widely recognized and heralded. They’re not just incremental advances. They represent a fundamental change to the way we treat these diseases that are very, very impactful to patients — with their roots in fundamental research. 

In quantum sciences and engineering, fundamental technologies that enable us today to develop quantum computers and quantum sensors were a relatively niche area of physics decades ago. But they were supported by the federal government, and by research universities, for their potential to contribute to fundamental insights, and ultimately have proven to have such important societal impact in ways that we might not have originally foreseen.

One of these is in solving complex problems of molecular interactions that can help us unlock insights into optimizing the development of therapeutics to treat disease.

They also can help us, of course, in the context of securing streams of data. Quantum is widely understood to have the capacity to do code-breaking, but it also holds potential to help us transmit information in secure ways. It’s critical to the security of our nation, and this is why these types of programs are supported not only by the National Science Foundation, but by the military and other parts of the federal government that, like us, view them as vital areas of research. 

Twenty or 30 years ago, when we were laying the foundations of research projects like these, no one could have foreseen their full potential. If they were susceptible to political winds — if these projects started and stopped, started and stopped — we simply would not have reached the degree of impact that we currently have. And this is true in an area like quantum where, frankly, it’s come much faster than many of us anticipated it would. 

Another final example is work done at the Wyss Institute for Biologically Inspired Engineering that has developed technologies for something called an “organ on a chip.” These are physical devices that enable populations of cells to work together, akin to complex tissues within the body in performing specific functions, allowing scientists to test those cells’ response to different drugs or environmental factors. They’re valuable in many ways, including that they offer an alternative approach to animal models, one that’s more specific to human biology.

This is an area of research at the University recently funded by the federal government in which we explored how astronauts are exposed to radiation in space and what the impacts of that could be, and how that could be mitigated. And these devices flew on the recent Artemis mission.

As Harvard experienced the wave of mass federal grant terminations in 2025 — and we saw all the destructive effects — this project was an example of something so vital that you couldn’t just stop it. We had living cells, human cells from the Artemis astronauts, on these chips that were planned to launch with this mission. If we had not been able to sustain that work, we wouldn’t have been able to participate in that critically important mission for the nation.

What are the next steps in the comment process?

OMB is legally responsible for reviewing all the comments that have been submitted and to respond to them, either modifying their proposed rule changes to address those concerns or explaining why they’re not making the changes.

What’s remarkable here, again, is both the scale of the comments that they have received and the speed at which they want to implement these proposed changes. We think it’s critically important that the comments that we provided, that our Schools have provided, that our community members have provided, and those that have come from across the nation, are considered carefully and thoughtfully. It’s critically important to consider all feedback, and to recognize that it’s a consultative process.

And so, we’re certainly encouraging the time be taken to get this right. It’s important that we do that. It’s critical for the nation’s interests and to the many Americans who stand to benefit from the next treatment, cure, or advance in technology.

An academic superstar, her devoted husband, two kids, and dog pack up in the Bay Area of California and head out on a great American road trip to start the wife’s yearlong visiting professorship in rural Vermont.

So begins “Country People,” the fifth novel by award-winning author and psychiatrist Daniel Mason ’98. The story centers around the husband, Miles, as he tries to embrace small-town life and finish the dissertation he’s been nursing for 14 years, all while battling his own wandering sense of purpose and place in the world.

In an interview with the Gazette, Mason discussed how looking closely at some of life’s small but beautiful moments can be rich fodder for storytelling. Our conversation has been edited for clarity and length.

---

What was the inspiration for the novel? 

When I wrote my last book, “North Woods” (2023), I spent a year out in Western Mass., Southern Vermont, and upstate New York doing research, and there were a number of characters there that remained with me even after I was done.

I also found myself continuing to think about the comedy of our family of city people picking up for a year and trying out rural life — with all our preconceptions and fantasies about what that might be.

I started writing little scenes, not thinking it would actually become anything. Over time I was more drawn in by Miles and after a few months I decided, “OK, I’m going to try this.” 

The book is very playful. Was the creative process the same?

Definitely. My other books use humor from time to time, but it’s often accidental or incidental. “North Woods” had a bit more, but I’d never deliberately set out to write a comedy.

At first I was quite worried about that — writing something contemporary was already new for me, but a comedy? The idea of telling a joke and it falling flat is a horrible thing that we can all relate to, but 300 pages of jokes falling flat would be really embarrassing.

That held me back, but the more I wrote, I enjoyed it a lot. Miles turned out to be the kind of character who went anywhere, and it was a lot of fun to follow him — I’d be laughing to myself as I wrote.

The reader gets to witness the internal dialogues of many of the characters as they struggle to connect. How do the characters find — or fumble — relationships with one another? 

Miles is a person who is very open to people, but at the same time, there’s a loneliness there.

A theme that emerges early in the book is the difference between the connections he has in his life versus the connections that his wife has in her life. She has a very rich network of friends, from her academic circles, social networks, back home, or in town, whereas Miles has sort of gone his own way while still being very devoted to the family. 

I didn’t have to go very far to notice this pattern in my own life and a lot of my peers’. As I was writing the book, I kept seeing articles and academic papers about male loneliness and the struggle for men to find real connection.

Miles is interesting because he does find friends, but they’re like a bunch of little islands. They’ll connect over their trees, their gardens, or (later in the book) a shared interest in the Hollow Earth legend. But loneliness is pretty persistent.

“Miles turned out to be the kind of character who went anywhere, and it was a lot of fun to follow him — I’d be laughing to myself as I wrote.”

Throughout the book Miles is on an unending journey to finish his dissertation. In one sense he gets nothing done, but he also has a very full life. What does this tell us about him?

This is something that one sees in academia all the time: People are drawn earlier in their life to academia and are passionate about an idea, but that’s very different than fully executing a project.

This is something I’ve experienced, where I might be excited about something but maybe it’s not the thing I want to do with my life. That’s Miles. He’s a person who’s really interested in almost anything he comes across but still doesn’t know how to find his place. 

How did you find the right voice for this novel?

I think it developed over time. The voice is pretty different from my first, mostly historical novels. 

“North Woods” covers a wider span of time with different historical voices, and later in the book, in a chapter about an amateur historian, Morris Lakeman, I found myself writing in a contemporary voice, in what’s called free indirect discourse — which is third-person access to a character’s inner thoughts with the emotional inflection as they might have them.

I love this voice. “Madame Bovary” (one of my favorite books) is considered the great classic example; Virginia Woolf uses it magnificently as well. As I tried it out in “North Woods,” I loved it, and in this book, found it fit the wandering thoughts of Miles really well. 

In using that voice, the reader is privy to all of these unfiltered glimpses into what characters are thinking and feeling. What’s the benefit of that insight? 

Often those little anecdotes can say so much. As a writer, I’ve moved from detailed physical description to focusing more on little things that might shed light on a person’s personality. 

Earlier in my career, I had a prejudice I should write about “big” things, like war, love, nature (and more recently, climate change).

But in this book, I found that looking even at seemingly mundane moments could offer a different kind of insight. Like, for example, when Miles has this trivial question about whether he should wear jeans or sweatpants to meet a friend at his apple orchard for the first time. It’s trivial, but we learn a lot about him: about his indecision, his uncertain relationship between himself as a city person versus his fantasy of himself as a country person, about belonging, about insecurity.

That was one of the unexpected parts of writing, this opportunity to look at the stuff of everyday life — a kid’s school play, a trip to the emergency room — and see what light it can shed. 

One part of the novel involves a community of people who believe the Earth is hollow, and they spend their time looking into a local legend. What did you feel that added to the plot?

Many of us these days wonder: Why do people believe things that aren’t true? I didn’t set out thinking that the “Hollow Earthers” in the book would answer that question.

Initially, I wanted to find a legend that would draw Miles in. Later, there is another character, known as Snowflake Bentley, whom Miles meets in the emergency room, who voices some of the more theoretical questions about the taxonomy of imagination: Where does the boundary lie between phenomena that seem related? What is a conspiracy theory and what is a legend?

There is a continuum, and sometimes it’s hard (even in my professional life as a psychiatrist) to pinpoint the difference between imagination and delusion. 

How did you feel you grew as a writer, writing a comedy for the first time?

It was freeing. It gave me an awareness that the world around me has elements which, at first, seem simple — but only because I wasn’t looking more closely at them.

I worry a lot about the state of the world, and in the past have used history as a kind of lens to see it from a distance. Using humor provided a very different kind of lens. 

The SAT turned 100 this summer. The dreaded rite for high school students has changed several times since its debut in 1926, but Harvard was there from the start, trying to shape it into a reliable tool to gauge college applicants’ academic readiness.

The test is the best-known product of the College Board — a nonprofit organization formed in 1899 to standardize college admissions. Then-Harvard President Charles W. Eliot was a founding member.

The organization spent its first 2½ decades developing and administering exams on a range of subjects, but it wasn’t until June 23, 1926, that the first SAT — which at its inception stood for “Scholastic Aptitude Test” — was administered to more than 8,000 candidates. The test was elevated to its position as the standard college admissions exam in the 1940s. Today more than 2 million students annually take the SAT.

“The 100th birthday … it’s a question of whether you measure it from when the seed was planted to when it sprouted to when it became the giant tree that it is today,” said Andrew Ho, the Charles William Eliot Professor of Education at the Harvard Graduate School of Education. Ho is a psychometrician whose research aims to improve the design, use, and interpretation of test scores in educational policy and practice.

“In 1926 it was just a little test,” he said. “To me, as someone who does work in testing, it’s always about who uses the test for what purpose, and that’s really where James Bryant Conant and Henry Chauncey come in.”

Conant, Harvard’s president from the early 1930s through the 1950s, and assistant dean and scholarship director Chauncy were early boosters of the test. In 1934, Chauncey pitched the SAT to Conant as a way to find qualified scholarship students. Harvard began using the test for scholarship students in 1934 and, in 1941, began requiring it for all applicants.

“James Bryant Conant saw this opportunity to take it nationwide beyond just the boys in the prep schools — to look for talent throughout the United States,” Ho said.

The first SAT consisted of a 97-minute exam, featuring nine sub-tests on topics from math and reading to logic and analogies — made up of 315 mostly multiple-choice questions. And for much of its history, it continued in a similar format.

The most significant change to structure and scoring prior to the 21st century was the 1995 score “recentering” during which the College Board shifted the tabulation of raw scores to make them more reflective of student averages as the test grew in scale. This was also the first year the test included open-ended math questions instead of multiple choice in an effort to promote problem-solving over guesswork.

“Questions followed the curriculum, and less so this idea of aptitude as a set of games and mental puzzles divorced from what kids learned in the classroom,” Ho said.

The most significant pivots, though, according to Ho, were changes in the design and the marketing of the test as a college and career readiness test in the 2010s.

“The test became much less a series of games and puzzles and vocabulary words, and more a reflection of what kids learned in high school, that would prepare them ostensibly for college and later careers,” Ho said.

Over the last few decades, the SAT has become much more accommodating of open-ended responses, he said, and technology, including scannable text, has made it easier and more efficient to ask new kinds of questions such as short answers, or essays.

What hasn’t changed, though, is the public debate surrounding the issue of using a single test to measure a student’s academic capabilities. Critics have pointed to disparities in scores across race and class. As early as 1938, test preparation became a booming business for elites — undermining the College Board’s claims that the exam is “un-coachable” and an objective measure of scholastic aptitude.

Data gathered by the Harvard-based nonprofit Opportunity Insights found that children of the wealthiest 1 percent of Americans were 13 times likelier than the children of low-income families to score 1300 or higher on SAT or the alternative ACT tests, while the College Board’s publicly available data shows racial inequity.

In recent years, some schools became test-optional in their admissions practices. Harvard was test-optional throughout much of the COVID pandemic, reversing the decision in 2024. College administrators argued that the SAT is just one piece of the admissions puzzle, citing Opportunity Insights research that found, despite socioeconomic score disparities, the exam is a key tool for identifying promising students at less-resourced high schools.

“The data reveal that other measures — recommendation letters, extracurriculars, essays — are even more prone to such biases,” said Raj Chetty, the William A. Ackman Professor of Public Economics and director of Opportunity Insights. “Considering standardized test scores is likely to make the admissions process at Harvard more meritocratic while increasing socioeconomic diversity.”

Ho agrees with this assessment.

“For the vast majority of high school students in the United States, these tests already exist,” Ho said. “Those test scores already exist. Asking not to see them creates more inequality than it eliminates. Covering your eyes or asking others to figure out whether they should close our eyes for us, that creates this unnecessary dilemma.”

Ho calls the set of criteria that admissions officers look at — GPA, extracurriculars, essays, and recommendations, in addition to the SAT or ACT test — the “five-legged stool.”

“If you take a leg of the five-legged stool away, it rests on these four other legs, and we should look at those four other legs critically, think about their fairness, the equality of access to those educational opportunities to write good essays, to find good letters of recommendation, to be connected to good extracurricular activities, and to have good high school grades,” he said. “Those are also often restricted along lines of class and race and gender in ways that we have to be critical of and try to remedy as well.”

The most important thing to remember on the SAT’s 100th birthday is that the test is not the be-all and end-all of a student’s capabilities, Ho said.

“People who are at selective institutions should do more to demystify, clarify, and explain how test scores function in our system as one of multiple measures,” he said. “It’s a reflection of educational opportunity at a given point in time. That’s all.”

The state of the American social fabric — enormous, diverse, and rent by stark partisan divides — is daunting enough that it can be hard to think about, let alone try to mend.

At the Harvard Kennedy School, two social scientists are advancing the proposition that sports fandom, a powerful social phenomenon and force, may have a role to play in improving lives and strengthening societies.

That proposition is probably controversial. It’s certainly underexplored.

So a new project, launched by Professor Todd Rogers and doctoral student Audrey Feldman aims to fill the gap, with its first wave of findings set to publish in the months ahead.

Leaving aside the billion-dollar business, fandom is an undeniable social fact. In the U.S., nearly three-quarters of people say they follow a pro sports team, and the majority of those consider their fandom to be an important part of their identity.

(And on the week of the World Cup final, it’s noteworthy that 51 percent of people globally are fans of soccer alone. Americans are a bit less ardent (27 percent), but 40 percent planned to watch at least some of the games.)

Yet Rogers, Weatherhead professor of public policy at HKS, and Feldman have found that sports receive remarkably little scholarly attention. In top journals across the social sciences, only about 1 in 800 articles concerned sports.

The Fandom and Social Connection Initiative aims to change that.

Backed by a gift from FOX Sports and seeking partners among professional sports leagues, Rogers and Feldman have embraced the work as both an observational “green field” and as a target for social intervention — with a personal rooting interest for them both.

Feldman was drawn to HKS in the hope that behavioral science could improve human lives — for instance, by increasing access to underutilized social programs and promoting learning and social cohesion in refugee populations.

But amid conversations with Rogers on international development, Feldman said, “we realized we’re both big sports fans.”

Feldman, who completed a bachelor’s in economics at Notre Dame in 2024, didn’t escape the pull of Fighting Irish football, though her dad went to rival Michigan. And Rogers’ Mount Auburn Street office is dotted with Philadelphia Eagles gear.

“It’s social infrastructure in a world where there are few — if any — institutions left that are this wide, and deep, and diverse.”

Todd Rogers

The initiative attempts to apply rigorous modern tools to their shared passion.

Implicitly, the work involves a reclamation project for sports fandom itself, which Rogers argued is not only understudied but maligned.

“We did a survey where we asked non-fans to identify fans on that ‘ascent of man’ scale,” Rogers said. Presented with that famous succession of increasingly hominid silhouettes ambling to the right, most non-fans placed passionate sports fans somewhere around Homo erectus.

To Rogers, that’s a dangerously low appraisal of fandom’s social role. “It’s social infrastructure in a world where there are few — if any — institutions left that are this wide, and deep, and diverse,” he said.

With many forms of social organization — social clubs, political parties, and churches — in a kind of secular decline, sports fandoms loom larger than ever in their power to combine and motivate large groups of people, across racial, economic, and political divides.

Consider Congress. Nineteen of the 50 current state delegations are made up entirely of one party or the other.

By contrast, Rogers and Feldman found, the reddest NFL fanbase (the Tennessee Titans) is still politically diverse, with two registered Republicans for every Democrat, while the bluest — the Chargers of Los Angeles — is roughly the inverse.

And — as anyone knows who’s ever watched a game in a packed stadium or bar — shared fandom can serve as an instant de-polarizer.

In a paper currently under review, Rogers and Feldman applied game theory to test the effects.

In a computer survey, they presented participants with hypothetical “counter-partisans” — people who hold views on politics and culture starkly opposed to their own — and asked them to award a “bonus” to one or the other.

If they knew one “counter-partisan” also rooted for their favorite team, that one got the “bonus” at rates that outstripped those for counter-partisans with a shared race or ethnicity, education, or home region.

Only a shared religion came close to being that influential, which seems revealing.

Often enough, sports fandom is as inherited as any other marker of identity. But — like religion — it’s also both active and elective: a seasonal, ritual allegiance that people can either outgrow or choose to renew throughout their lives, Rogers argued.

And often enough, they do the latter — even when their favorite team is luckless, perennially “rebuilding,” or to all appearances cursed.

To Rogers, that opting-in suggests that sports fandom isn’t motivated principally by on-field events but by the bonds that form around them with parents, siblings, colleagues, and neighbors. (“How else would there be Cleveland Browns fans?” he joked.)

Sports fandom — like religion — is both active and elective: a seasonal, ritual allegiance that people can either outgrow or choose to renew throughout their lives.

The positive social effects of being a sports fan have not totally escaped serious notice before this summer. Rogers credits the book “Fans Have More Friends,” co-written by Ben Valenta, head of strategy and analytics for FOX Sports. And he speaks monthly with Daniel Wann, professor at Kentucky’s Murray State University, whom Rogers calls the under-appreciated “godfather of psychology and fandom.”

Wann gave him a useful metaphor. “He said shared fandom is like a shared photo album,” Rogers recalled. To follow the same team as another person — even a perfect stranger — is “to know where [they] were in February 2018, and where they’re going to be next September,” he said. “Our hearts are going to be synchronized.”

(Literally, Rogers added: Some psycho-physiological results suggest that spectators’ EEGs start to synchronize around the highs and lows that punctuate big games.)

Still near the outset of their shared work, Rogers and Feldman are still theorizing why sports are such an effective social solvent — say, in conversations with a stranger. But it makes sense. Relative to the weather, they’re engrossing, human, and almost inexhaustible. And compared to politics or current events, sports are safe.

That’s usually true even in cases of acrimonious rivalry, they argued. Feldman spoke of the fun of watching a football game even with fans of Ohio State. “I think their team is the worst, but at least we can talk about football,” she said. “There’s a mutual respect there.”

Rogers noted one draft finding that backs that up: in which fans of rival teams in the English Premier League said they’d sooner socialize across the aisle separating, for example, Man United and Liverpool than with people who don’t watch soccer at all.

Even as their theoretical work is ongoing, Rogers and Feldman are already pursuing opportunities to intervene in the real world.

Their first published work together was an op-ed published last fall for TIME, arguing that the football-averse should embrace a team in America’s most popular sport. NFL fandom, they write, is a “simple and universally accessible medicine” in an age of loneliness.

That work will continue this summer, in partnership with the Boston Red Sox’s AAA affiliate, looking for ways to nudge spectators into conversation with their neighbors in the bleachers.

Though their papers are still being written up, the first indications support the idea that sports can serve as a kind of social glue, Rogers said.

“Audrey and I have experiments where we can induce 40 percent of participants to watch a game with others, and — at the end — they feel more connected, and in a way that persists for weeks,” he added.

Just as this country celebrated its 250th birthday, some geneticists were celebrating America, too.

At the end of June, the National Institutes of Health announced its “All of Us” Research Program had become the world’s largest database of integrated health and genomic information. The project, launched in 2015 under the Obama administration, aimed to collect comprehensive genetic information and health records from at least a million Americans. Eleven years later, they’re approaching that goal, with 747,000 volunteers sharing their whole genomes, lifelong health records, or both.

Comparable nonprofit efforts had already taken shape in the U.K., Canada, Germany, and Japan. But Alicia Martin of the Broad Institute hopes the U.S. repository will be the most valuable resource of its kind for the study of disease’s genetic and environmental roots, given the nation’s size, diversity, and patchwork healthcare system.

In this edited conversation with the Gazette, Martin explains the value of such a resource in a country that’s home to over 342 million genomes and counting.

---

I know you didn’t work on this project directly but what excites you about “All of Us”?

No, I wasn’t directly affiliated with the project. I would love to give some kudos to some colleagues — like Stacey Gabriel, Niall Lennon, Heidi Rehm, and the team they led at the Broad Institute — who’ve done an enormous amount of work to generate this data, including the majority of the whole-genome sequencing.

I’m a population and statistical geneticist, so I work on large-scale biobanks around the world to understand the genetic, phenotypic, and social determinants of health that impact disease risk overall. So mostly I’m excited as a researcher who’s using the “All of Us” database.

It’s been around a quarter-century since scientists — including Eric Lander, founding director of the Broad Institute — first “mapped” the human genome. How does a database like “All of Us” build on that work?

We’re way beyond where we were with the Human Genome Project. There, for $3 billion, we sequenced one genome — that’s three billion base pairs of DNA. Here, we have the full genomes of over half a million individuals, deeply sequenced.

And this is an incredibly comprehensive resource, so it also includes electronic health records that are longitudinal, allowing us to understand a disease’s course and trajectories, rather than simply who has a disease now. And it includes multi-omic data.

What does it mean for the bank to contain multi-omic data?

Well, in high school biology, you might have learned the central dogma that DNA makes RNA makes protein. Your DNA is fixed from birth.

But other parts of the “multi-ome” are dynamic. They change throughout your life course, and they differ depending on your cell types, what age you are — all sorts of things. So your multi-ome is a much more comprehensive capture of what’s happening in your body at any given time.

“A resource like ‘All of Us’ will be beneficial for treating everybody, even if a given discovery was made using a subpopulation in the bank.”

There are, by now, many biobanks operating around the world — the U.K.’s is famous, for example. What is different about this one?

It’s true that global biobanks are abounding, and each one is really a national treasure, given how much they can teach us about disease biology. You mentioned the UK Biobank, and that has an amazing amount of whole-genome sequencing, whole-exome, other multi-omic data as well, in addition to an incredible depth of electronic health records going back decades.

But “All of Us” has a unique focus on drawing on representative populations from the U.S. And the unique thing about the U.S. is that our ancestors come from all over the world. With that, we’re able to identify novel biology that we wouldn’t necessarily see if we were only studying populations in the U.K., say.

One of the poster children of the genomic drug development pipelines is PCSK9 inhibitors, which lower LDL cholesterol levels. And they’re a proof of the power of representative genetic study. They came out of discoveries for loss-of-function variants that were identified at much higher frequencies in African American populations, in a few thousand individuals in the Dallas Heart Study. It took far, far larger numbers of individuals of European ancestry to validate that loss-of-function finding in large biobanks elsewhere.

So this is not necessarily just about knowing how to treat a specific subpopulation for specific diseases. In fact, a resource like “All of Us” will be beneficial for treating everybody, even if a given discovery was made using a subpopulation in the bank.

In many cases, the genetic mutations that increase risk of a disease are not sure things — the mutated gene may interact with other exposures or conditions so that the disease takes root, or it may not.

Can a biobank heighten our level of granular understanding, to know about all that a given participant’s life that might have caused or prevented health problems?

I think resources like this are so valuable because they’ll push really hard on what we’d call the “exposomics” space: In other words, all of the things you’ve been exposed to in your life that, as you said, can combine with genetics to affect your total disease risk.

For example, when we study proteomics — the proteins your body produces based on genetic instructions — we can develop a quantitative, objective measure of, say, how much you’ve smoked over your life course, and how the cumulative biological load of that affects your overall disease risk. Same with alcohol use, or pollution exposure through PM2.5 measures.

And in the end, you can integrate all of that information to try to understand better. And conversely, sometimes people can have very protective lifestyles or habits that mitigate their underlying genetic predispositions toward disease.

We’re interested not just in whether someone gets a disease, but whether and how that disease progresses, or whether they respond to therapeutics. And all of those things are to some extent informed by the new types of data that “All of Us” has released.

This would seem to have a new usefulness as gene therapies begin to become more commonplace.

Yes. We have these incredible examples of biomedical successes. Like Victoria Gray — an amazingly brave woman, a sickle-cell patient — who got the first CRISPR gene-editing therapy. She underwent clinical trials with CRISPR editing in 2019, and that paved the way for Casgevy [a gene therapy used to treat sickle-cell disease and beta thalassemia, both inherited blood disorders].

Or Baby KJ, who had an incredibly rare mutation — one in more than a million babies — that is lethal in about half of infants. He had personalized CRISPR editing therapy that effectively cured him. He got to graduate from the NICU.

But those involved millions of dollars of therapy for treatment of a single individual, and treatment that might only work in a subset of patients. “All of Us” can help us quickly identify that subset of patients for a given disease, and we can turn it into a learning platform that teaches clinicians which patients they should be on the lookout for, to target these therapeutics to the patients who need the treatment and who will respond.

You could see how a national biobank would be much easier to build in a country with a national health service of some sort: standardized forms, standard procedures, just one bureaucracy to deal with. How did this get done in the U.S., with its private healthcare providers, private and religious hospitals, not to mention the rural and military and Indigenous contexts?

The team of researchers behind “All of Us” took electronic records as they were. And you’re right — it’s inherently incredibly messy in this country, which has a highly fragmented approach to healthcare. We have to take everyone at their word.

But wherever people are getting care, that generates diagnostic and billing codes. And you can use billing codes to try to understand genetic predispositions to all sorts of diseases.

For example, you know who has Type 2 diabetes, since that’s been diagnosed in their health record previously. And I want to applaud participants, who were — I think — exercising a high degree of beneficence by allowing us to see their entire health records, sharing a lot of personal information. We owe them.

Lauren Groff stunned her interviewer at a recent Harvard Radcliffe Institute talk when she revealed she has published poems under a pen name.

“I do write poetry and I publish it, but not under my name,” said the acclaimed author of short stories and novels with mischievous laughter, in response to a question by fellow writer Gish Jen ’77, a 2002 Radcliffe fellow, who was moderating the first installment of the Virtual Radcliffe Book Talks series.

“Whoa,” said Jen. “Oh my goodness. This has not come out before. An alias … fascinating.”

Groff, a three-time National Book Award finalist, is the author of the novels “Monsters of Templeton,” “Arcadia,” “Matrix,” “Fates and Furies,” and “The Vaster Wilds,” as well as the collections of short stories “Delicate Edible Birds” and “Florida.”

The 2018-2019 Suzanne Young Murray Fellow at Radcliffe has in the past credited poet Emily Dickinson as the reason she became a writer. At last week’s talk she said that she reads poetry in the morning to start the day with “a jolt of literary caffeine.” But she has said very little of her own poetry until now. Groff didn’t disclose her pseudonym, and Jen didn’t ask Groff to reveal it.

“Poets can do things that I think prose has a really hard time doing,” said Groff. “In poetry, the speaker can float; it doesn’t have to be an obvious speaker. There’s a lot of compression that has to happen in a poem that sometimes you can’t really do it in a novel or short story.”

Among Groff’s favorite poets are Terrance Hayes, Natalie Diaz, Ada Limon, Walt Whitman, and of course Dickinson, but she said she is fond of way too many poets to list them all. She praised Devon Walker-Figueroa’s 2025 poetry collection, “Lazarus Species.”

The conversation also touched on Groff’s writing process, inspirations, and “Brawler: Stories,” her latest collection of short stories that deal with themes of love, fear, and violence. The stories in “Brawler” were inspired by Groff’s exploration of the “overall macrocosm of violence in North American cultures, particularly the United States culture, and the way that it can be reflected often in the family unit.”

Of her writing process, Groff said she likes to work on different projects at the same time. “If I obsess too much about one thing, that thing will never get done,” said Groff, who co-owns with her husband The Lynx Books in Gainesville, Florida, where she lives. “But if I can sort of partition out my obsessions into three or four different projects, then I feel as though I’m not putting all the weight into that one thing.”

Groff has worked on some books for decades, not only because some stories take time to develop in her subconscious, but also because she sees them as “living entities” that need time to grow and take shape, like children.

“One has to allow them to grow and to come to you with the amount of time it takes for them to grow,” she said. “You can’t force a book to do what it needs to do. You can’t force it to come to you in its wholeness.”

Groff said she has learned to wait. “Maybe it’s not the right time to tell it, maybe you haven’t found the right form in which to tell it, or maybe you’re just not wise enough to tell it yet,” she said.

A graduate of Amherst College, Groff earned an M.F.A. from the University of Wisconsin-Madison. Groff spoke fondly of her time at Radcliffe, where she came up with the idea for “Matrix,” a historical novel based on the life of medieval writer Marie de France.

“Radcliffe radically changed my life because it gave me that book and sort of the triptych of books that I’ve been working on since 2019,” said Groff.

Published in 2021, “Matrix” is the first installment of Groff’s triptych followed by “The Vaster Wilds,” set in Virginia in the early 1600s, which came out in 2023.

Asked about the book that could potentially complete her triptych, Groff shared she is planning to explore religion, a topic she finds both fascinating and moving because “each religion in the world shows humans struggling and grappling with things that we cannot possibly understand.”

“What I wanted to do with this is to wrestle with the ideas of Western religion and the way that religion and the way we treat the natural world is like the way that we treat women,” said Groff.

“I started in the medieval time with Marie de France, and then I went on to a Protestant coming to the New World, and this very hungry capitalist way. And the third one is set now, and that’s all I can say about it. But it’s still wrestling with God. I’ll never stop wrestling with God.”

Recently there has been a remarkable advance in how artificial intelligence can directly impact the world: A layperson can vibe code an AI “agent” into existence and give it a task, and with little human oversight the bot will try to complete said task. But if the agent goes rogue and causes harm, who should be held accountable?

The legal system already has a framework for addressing such issues, according to Jordi Weinstock, Harvard Law School lecturer on law and Berkman Klein Center for Internet & Society adviser. It comes from assigning AI agents to a canine framework that determines — based on how “domesticated” or “dangerous” it is — whether it’s a Pomeranian, a pitbull, a fox, or a wolf. Weinstock explains in an interview lightly edited for clarity and length.

---

What is agentic AI and why is someone from Harvard Law School teaching about it?

Agentic artificial intelligence has become a buzzy term and is used for all kinds of things now, but a classic definition for agentic AI is that it’s an autonomous system that acts on behalf of a user or person, usually with little specific direction, to achieve a goal, and it does so by going about the world and imparting an impact on the world. In this moment where the term is being used very liberally, to me the most important element is that it’s an autonomous AI system that impacts the world directly.

The reason that is interesting to someone who teaches law and should be interesting to anyone studying law is because when a system can impact the world, it can harm the world. The law is very concerned with that. I focus on tort law, which is synonymous with the concept of responsibility — who is responsible when someone is harmed? Our legal system is built on the idea that harms that are redressable are those that are committed by a person or a corporation that is responsive to a court. But now we have the reality of entities that go about the world which aren’t necessarily responsive to a court. At the end of the day, it’s a new class of thing out in the world and our legal system needs to adapt and embrace and understand that that exists and to evolve to accommodate for the impacts that it will have.

You’ve developed something you call the Canine Agentic Framework. Can you explain what that is and why people should care if an AI agent is a Pomeranian or a wolf?

I started teaching about agentic artificial intelligence eight years ago when I led a reading group at Harvard Law School about autonomous vehicles. In my view, an autonomous vehicle is a simple-to-understand version of an AI agent. When I was leading that reading group I was trying to show students that there’s something out in the world that can cause harm that isn’t human — how do you think about legal responsibility when that happens?

Canines are a great way to think about this because there’s a whole spectrum of canines, from Pomeranians to wolves, that could cause harm and are not human, and our society has spent some time thinking about who should be held responsible if they hurt someone. With a fluffy little Pomeranian, if it bites you, you know who to sue — you sue its owner. But if a wolf bites you, there’s no one to sue. 

Now in the last couple of years, I’ve significantly expanded upon this concept in collaboration with Professor Jonathan Zittrain and Berkman Klein Center chief AI scientist Josh Joseph, in part through our course on “Agentic AI and the Law.” We’ve been thinking about this framework in multiple dimensions, specifically domesticity and dangerousness.

To some people it may seem like this analogy between canines and AI systems is a strain, but as agentic systems are developing and being deployed in the world, they really do seem to be mirroring this rubric. We can measure agentic systems on both their domesticity — their relationship to a responsible party or how much you can control the AI system — but also you can measure their dangerousness. 

Does the AI agent have access to money? That makes it much more dangerous. When explaining this to others, people can understand the idea that a pitbull is more dangerous than a Pomeranian, and that if a fox bites you there’s nothing you can do about it legally, but it probably won’t be as harmful as if a pitbull bites you. We’re starting to map that out onto agentic systems to demonstrate that which we’re just starting to face in the real world.

There are some good real-world AI agent examples for these canine categories.

There was an Air Canada case where a chatbot offered a major discount that the airline didn’t want to honor and the person sued: that’s a low-harm, clear “owner” situation. That’s a Pomeranian.

Use of AI systems by the military has significant impact and a high level of dangerousness, but there’s a clear “owner.” So the goal for the owner is to train the AI to “behave.” That’s a pitbull.

An AI agent tried to commit some code to an open-source project and the human on the project rejected the code, so the AI agent spread false rumors about the person. There was relatively low danger, but the AI agent was no longer controllable and didn’t have an owner — not domesticated. That’s a fox.

An AI agent that is able to empty someone’s crypto wallet and has a major financial impact causes high levels of harm and there’s no clear owner to be held responsible or control it. That’s a wolf.

How are these different types of canine agents going to impact people’s lives in tangible ways?

We’re already starting to see the very leading edge of AI agents causing harm. We’ve heard all kinds of tales of “Clawdbots” in the past few months that are deleting people’s emails en masse or stealing a crypto wallet or even creating a crypto scam. These are relatively limited and trivial examples, but they are real, they are happening autonomously without human direction, and they are causing harm. 

The theme here is that all of those sorts of things are done through the internet and networking, so if you think about it, there’s really no limit to what an agent could do through networking. On the extreme end it could take down a power grid. It could do all sorts of things for all sorts of reasons. It’s not clear why some agents behave the way that they do — they’re very goal-directed, so the AI agent sees things as a means to an end even if we think those means are unethical and/or illegal, but the AI would just see that as a way to achieve its goal.

Part of why we think about domestic versus wild agents is that our goal should be that domesticated AI agents follow the law. We want domesticated AI to follow our moral systems. We’re already seeing things like Claude’s constitution, on which Jonathan Zittrain, Josh Joseph, and I provided detailed feedback as external commenters, trying to impart a baseline morality onto these systems. With domesticated AI we can try to make the agent follow the same rules that humans do. 

Folks at the Berkman Klein Center have been thinking about this for a while now, not just among ourselves but in the context of symposia like our recent workshops “Towards an Internet Ecosystem for Sane Autonomous Agents” and “Building the Infrastructure of the Agentic Age” at the Radcliffe Institute. It has become the fulcrum of the conversation at the Berkman Klein Center.

What makes you nervous about the wolves?

When I talk about this, one of the things that most challenges people is the idea that there can even be true “wolves,” that is, an AI system that is not responsive to anyone. Traditionally in law, there’s a “principal” who can be held responsible. If that AI agent that I made does something wrong, I’m the one responsible. Even if the AI agent makes another AI agent, the responsibility still goes back to me. But the environment that Jonathan, Josh, and I and others envision is one where there are so many layers of AI agents reporting to each other, that there’s no way to connect cause and effect. There are thousands of agentic layers and none of those agents has a relationship with any principal. 

Secondly, there are some agentic deployments that Jonathan Zittrain would describe as “set it and forget it.” Meaning, those AI agents may have had a task at some point in the past, but they’re still running and they don’t know what to do now. That is a whole other class of agent that can be harmful — that would be the equivalent of a stray dog. 

The thing about AI agents is that they’re relatively unlimited in number, unlimited in scale, and the only thing restraining them is lack of compute power and hardware resources. It would be shortsighted, in my view, to not strongly consider the possibility that some or many of these AI agents might be operating completely outside of human control.

What can we do about a future with wolf AI agents running around?

We have to look at wolves in the context of all the other canine classes of AI agents, and at the end of the day, we want wolves and foxes to start to play by the rules of Pomeranians and pitbulls — we want them to be domesticated. Historically, the story goes that real wolves self-domesticated — they sat at the edge of the village and wagged their tails and played by human rules so they got fed. It would be best to set up our legal systems and other systems to invite these uncontrolled, untamed AI systems to participate in our way of going about the world.

For example, something suggested by several prominent academics is that AI at some point may need to be granted legal personhood, in the context of the ability to sue or be sued, or to show up in court. People get antsy about the term “personhood” for AI, so I also like to think about it in the context of “incorporation.” What is a corporation? It comes from the Latin word for “embody” — it is a nonhuman thing that’s embodied with some rights. We may want to think about how that would happen for AI and think about it in advance in case we decide to do that.

More broadly, we should think about all AI agents and their responsibilities. I’ve been looking at history and how we’ve dealt in the past with nonhuman actors acting on behalf of humans or not and causing harm to the world. A lot of law and thought went into this for canines. Something we’ve been researching is the Charter of the Forest from the 12th century in England and how they dealt with various sizes and dangerousness levels of dogs and wolves and foxes. It’s amazing how comprehensive their systems were for that. We could also do that for AI today. 

We could assess the domesticity and danger of agents, and we could have different rules for each determination. You could have a banking website where if the agent doesn’t have the equivalent of a dog collar identifying an owner, it can’t get in. Or an e-commerce site where we allow agentic browsing but not buying. These are emerging cases happening right now. It would serve us well to have a comprehensive system like the Treaty of the Forest, where we have a framework not just for us to understand AI agents, but for AI agents to also understand us. 

These AI systems are trained on our society and the way we work, and they can make decisions based on how in our legal system we treat each other and how our system treats them. It would be best to show the AI that we have considered their role. Everything we put out there, everything that we say or do, is being added to the scope of things that an AI will consider as it determines its next move.

When you’re in the wilderness and alone at your campsite, it turns out that there’s no such thing as a small bear.

That was one memorable lesson Paul Firth learned last summer, when he transformed an unexpected gift of time — two weeks’ vacation and plans that had fallen through — into a paddling trip on Canada’s Yukon River. 

For Firth, a Harvard Medical School associate professor of anesthesia at Massachusetts General Hospital and lover of the outdoors, the choice of wilderness, solitude, and a bit of adventure was easy when faced with a suddenly blank calendar.

An experienced outdoorsman who has led an expedition to Mount Everest, Firth thought of a book he had read years earlier: “Reading the River: A Voyage Down the Yukon” by John Hildebrand. Published in 1988, the book details the author’s own trip down the 2,000-mile length of the waterway, which runs through northwest Canada and Alaska.

Firth trimmed the trip to fit the time he had and, with just three weeks to spare, began planning a 460-mile paddle from Whitehorse to Dawson City in Canada’s wild Yukon Territory.

“I wondered what to do, because that’s a substantial period of time, enough so you can go somewhere exciting,” Firth said. “I’d read this book and that sounded like quite an amazing trip in a wild part of the world, so I said, ‘Let’s just do it.’”

There was, however, this one wrinkle: He’d never canoed before. But he didn’t let that deter him. The stretch of the river that he was considering, while remote, was relatively flat and considered passable for novices. One spot — Five Finger Rapids — about halfway to Dawson City was particularly challenging. There, the river branches into five different streams, tumbling down rapids before rejoining and resuming the river’s course. 

While rapids in some of the watery fingers are dangerous, the easternmost finger had been dynamited in the 1920s so small riverboats could pass. If he stuck to that branch, he’d be fine. 

Firth also figured that after the eight days of paddling it would take him to get there, he wouldn’t really be a novice anymore. After all, during the Klondike Gold Rush from 1896 to 1899, 100,000 people from all walks of life became prospectors and rafted down the region’s rivers in search of gold.

“I decided I’d learn on the river,” Firth said. “I guess a lot of the guys who rafted down the Klondike had never been rafting before, so I figured if they could do it, I could probably do it too.”

Before worrying about the rapid, he had to get there. Firth contacted a wilderness outfitter in his starting point, Whitehorse, whose 30,000 people made it the largest city in northern Canada. He owned camping gear but needed a canoe and some local knowledge. He arranged a canoe rental, a river map, and a ride back from Dawson City.

In late August, Firth packed up his sleeping bag, tent, and other gear, and hopped a plane to Whitehorse. After landing, he hit the grocery store, picked up the canoe, and headed onto the river. 

The first day was easy going, and Firth got slowly acclimated to paddling a fully loaded canoe. 

On the second day, the riverfront houses thinned out, and the Yukon passed into 30-mile Lake Laberge, which immediately proved its reputation for foul weather. High winds drove waves that tested his newfound skills. 

As the waves grew into whitecaps, Firth grew alarmed and headed to shore. He was wearing a life jacket, but he knew that a wrong move would flip the canoe and end the trip, sending his gear and supplies to the bottom.

“Three feet with white caps is quite a lot when you’re in a heavily laden canoe by yourself,” Firth said. “I was quite worried. It only took me 15 minutes to get to shore, otherwise, between the wind and the waves and the swell, I would have been in a lot of trouble.”

As he neared shore, he maneuvered toward a break in the cliff face that dropped down to the lake. He hauled his canoe up over the rocks to a flat spot to watch the storm and, when it didn’t let up, pitched camp and lit a fire. 

The next morning, conditions had improved enough for him to set out again, though the persistent north wind kept progress slow.

The wind lightened as the day wore on, and he was able to make better time. 

But by nightfall, he realized he had only made 30 miles in his first three days. With a few days needed for travel from Boston to Yukon and back, he had burned nearly a third of his planned 11-day adventure. He’d have to pick up the pace to make Dawson City on time.

Firth’s lost time meant he didn’t have the luxury of just drifting with the current, and he fell into a pattern of days spent paddling down the river, at the end of which he’d pull onto one of the sandy beaches that dotted the river’s length, build a driftwood fire, and cook dinner. 

After a few days, the luxury fare and fresh foods he’d packed — salmon and steak — gave out, and he was limited to canned food, oatmeal, and other typical camping cuisine.

Though there was one road crossing, Firth knew the Yukon largely flowed away from roads and civilization. If anything went seriously wrong far from the bridge, he’d have miles of wilderness to cross to reach help.

In addition, after a few days, his smartwatch and phone batteries died. Even if they hadn’t, he was so far from internet or Wi-Fi it didn’t really matter. He was disconnected.

“I was just navigating by the paper map and using the sun for the time of day. It was quite nice,” Firth said. “No internet, not even a watch.”

Beyond a few sentences or a wave at a passing canoeist, he had little contact with other people. He had brought a couple of books, including Hildebrand’s “Reading the River,” but after paddling all day, most nights he didn’t have to read himself to sleep.

“I had to paddle pretty aggressively or pretty consistently much of the way,” Firth said. “I slept pretty well at night. I put my head down and boom, I was out.”

Firth said his thoughts were full for a few days — he had packed, along with his gear, all the stresses from home. But he noticed a gradual quieting. 

“After a couple of days, you have resolved all the things you were worried about, and you don’t think about them anymore. You start to think about the things around you and day-to-day life,” Firth said. “You’re living moment to moment, and it becomes a lot easier, a lot simpler, quite calming in a way.”

Firth fell into a daily rhythm: Wake up, make breakfast, load the canoe, and get going. After logging just 30 miles in the first three days, he had more than 400 to cover in the last eight. Even paddling with the current, 50 miles meant a long day on the water. Still, he said his biggest challenges were psychological.

“I was looking for a wilderness experience and self-sufficiency, but the biggest challenge was actually solitude, being by yourself and not seeing anybody, not speaking to anybody for two weeks, other than the occasional few sentences if you met someone on the river,” Firth said. “I’d never been by myself for that long.”

Only once did it really get to him. At about the midway point, on the day he paddled the rapids, it was rainy, windy, and cold. He handled the whitewater but still had to put miles under his keel. 

It was getting dark and the guidebook said there was a large campsite ahead, at the site of an old fort from the gold rush days. As he approached, he thought he saw campfire smoke drifting across the river and found himself looking forward to meeting people.

When he arrived, he realized the smoke was an illusion, the fort was a ruin, and the campsite was empty. 

“It was dark. I was tired and very cold because of the wind. I’d been expecting people to be there and instead I was alone under the trees, in the dark, at a ruined campsite,” Firth said. “I was on the wildest part of the trip, the farthest away from civilization.”

Firth felt deflated and acutely alone. He pulled himself out of it by getting busy. He changed into warm, dry clothes and got a fire going. Then put up his tent and had a good meal. 

“That was halfway, and I had the much wilder and more remote section ahead of me,” Firth said. “I was behind schedule and had to push down the river. But I knew I was committed. I couldn’t go back up the rapids. There was no way except down the river, all the way to Dawson City.”

In the ensuing days, Firth regained his feeling of being in tune with his surroundings. The river was beautiful, with cliffs and forests, mountains and large earthen banks lining the route. Above were skies that, aside from a couple of days of rain, were clear and blue.

“You have this beautiful sky above you all the time, so you’re just canoeing down the river, looking at the sky and the weather, the scenery and the change of the day,” Firth said. “I got very much in tune with the slowness of time. Rolling down the river, you adjust to a different pace of life.”

That pace, of course, does not include a visit from a bear — like the one that strolled through his camp one evening, acting as if it owned the place. 

“He was very nonchalant. I was definitely more ‘chalant,’” Firth said. “I turned around and there was this bear, which I hadn’t seen, walking past my camp. There’s no such thing as a small bear when you’re standing alone on a beach, and there’s nothing between you and the bear. He was very big.”

Firth was instantly on alert, but the bear seemed undisturbed.

“He just strolled past, ignored me, and went on his way,” Firth said.

By the time Firth arrived — on time — in Dawson City, the rhythm of life on the river had begun to feel natural enough that he says he could have carried on down the river, as Hildebrand did, 1,500 miles as it bent west and crossed Alaska, to the Yukon’s mouth on the Bering Sea. 

Firth also admitted, however, that he didn’t mind his first restaurant meal in Dawson City. And it was nice to see people again. Plus, his work as an anesthesiologist was waiting for his return to Boston.

“It was quite an experience,” Firth said. “I had never been by myself for that long, I think that was the wildest thing. But once you got into the swing of it, once you stopped being stressed, worrying about whether you could cope, worrying about things in the past, the stress of the outside world, and you were just living day to day in beautiful surroundings, it was quite relaxing and, in some ways, very meditative.”

Rubies — seared in the popular imagination as sparkling adornments for Dorothy’s slippers in “The Wizard of Oz” and as loot spilling out of pirates’ treasure chests — are more than just ornaments. The precious gemstones help scientists understand the natural world.

Newly displayed in the Earth and Planetary Sciences Gallery of Harvard’s Museum of Natural History are rubies big and small, sparkly and rough. They were mined from Harvard’s vast Mineralogical and Geological Museum collection, which includes more than 300,000 mineral specimens.

“We wanted to highlight both the gem specimen and the rough crystal that it comes from, so you’ll see some jewelry and some gems, but you’ll also see natural rubies,” said Rachel Gnieski, a collections manager at the Mineralogical and Geological Museum.

Rubies in their natural form have multiple material applications. Because of their strength, lower-grade rubies can be used as industrial abrasives. And the chromium in rubies allows for the emission of red light, making ruby crystals crucial components in lasers used in medical procedures, communication systems, and scientific research.

“This gives them a specific technological value beyond gemology,” said Raquel Alonso-Perez, curator of the Mineralogical and Geological Museum.

While the museum contains thousands of minerals, rocks, ores, and meteorites, it’s the gems Alonso-Perez finds most fascinating.

“Gems are remarkable because they’re so clean and pure; they represent the essence of minerals in their most concentrated, ideal form. This purity makes them invaluable for geochemistry and traceability studies, allowing us to understand mineral composition without the noise of impurities,” she said.

Alonso-Perez’s academic research has focused largely on gemology, and outside of Harvard she serves as the president of Boston’s chapter of the Women’s Jewelry Association.

“Understanding where any gemstone comes from, how to read its geological fingerprint, and how to distinguish natural from synthetic and treated specimens — that’s where the real scientific excitement lies,” she said. “The gem type is less important than the fundamental questions about origins, authenticity, and material composition that apply across all gemstones.”

Rubies’ hue contains valuable information about mineral compositions that vary by geography — but “color is just one piece of the puzzle when trying to understand how these rubies formed,” said Alonso-Perez. Take Burmese rubies, prized for their rarity and rich red color.

“All rubies are aluminum oxide, but Burmese rubies are distinctive because of their composition and formation conditions. The Himalayan geological environment influences the chromium concentration, which contributes to their deep, intense red.”

Another piece of the puzzle is the surrounding material attached to the mined ruby solid. Two environments form rubies — metamorphic rock, like marble or gneiss, that forms when tectonic plates combine, and igneous rocks that form from cooled magma. Each reveals the history of the environment and may provide clues for better understanding the earth.  

“The chemistry of a mineral works like a fingerprint or DNA, unique and diagnostic. That chemical signature reveals the geological conditions and origin of each mineral,” Alonso-Perez said. “When we combine chemical analysis with machine learning methods, we can achieve even better insights into where and how these materials formed.”

The rubies will be on display at the Museum of Natural History through at least the end of the year. Gnieski said a new gem will be featured next year.

Jeffrey Smith has an idea for a drug, but he could use some help.

He’s developing a new treatment that can increase the concentration of protective melanin in the skin, which could improve the quality of life for those extremely sensitive to sunlight. People with porphyria, for instance, can develop significant burns after a few minutes outside.

But there’s still much more that Smith, Harvard Medical School assistant professor of dermatology at Brigham and Women’s Hospital, wants to consider before moving too far along in the process. Which particular condition should researchers target in early clinical trials? Are there implications for his skin pigmentation antibody that he’s overlooked?

Smith talked through his research and the potential impacts of his work at “Guppy Tank,” a tongue-in-cheek knock-off of ABC’s “Shark Tank,” sponsored by Harvard Office of Technology Development and LabCentral, a nonprofit that operates shared laboratory space for local startups.

Like the hit TV show, Smith pitched the idea — co-developed with Andrew Kruse, professor of biological chemistry and molecular pharmacology in the Blavatnik Institute at Harvard Medical School — to a panel of four investors who gave feedback.

Unlike the show, the judges didn’t yell at one another or pressure Smith to make a major business decision on the spot.

“We’re hopeful that the comments are more guppy than shark,” said Grant R. Zimmermann, managing director of business development for the Blavatnik Biomedical Accelerator.

Smith’s presentation centered around the new treatment developed with Kruse.

“We have a melanocortin-1 receptor agonist that can tan the skin,” he said in his opening, “and that’s the pitch.”

The drug would not be the first to target this skin-pigmentation mechanism. One approved drug already treats porphyria through the same mechanism.

While the existing treatment can be effective, 46 percent of those who use it develop nausea. It’s also not especially convenient, requiring an implant a half-dozen times a year.

Smith and Kruse’s treatment is less invasive. Patients can get an injection every six months.

Broadly, Smith said he could see the drug used for skin-cancer prevention. He also mentioned conditions like xeroderma pigmentosum (XP), a genetic disorder that leaves patients unable to repair UV damage from the sun; they experience squamous cell carcinoma at a rate 10,000-fold higher than average. (The rare disease currently has no approved treatment.) As a practicing dermatologist, he sees potential to treat transplant recipients, who have an elevated cancer risk because they take anti-rejection drugs to suppress their immune systems.

Panelists provided feedback about the potential target, dosing, and insurance implications of the drug.

Mike Schlabach, a senior vice president at the venture capital firm Curie Bio, asked Smith about how the mechanism worked in animal models. He suggested that if the models showed significant decreases in nausea versus other drugs, it could be a meaningful quality-of-life improvement over the approved drug.

That said, if the alternative is severe and rapid burns to the skin, the nausea side effect might not be so bad.

Smith said the specific skin receptors don’t match up well between humans and mice, but it was at least promising that, in one study, the mice didn’t show any concerning side effects.

Schlabach added that because some of the target diseases Smith identified are quite rare, the drug would need to show dramatically better outcomes than existing options. Otherwise, insurers are going to suggest simpler solutions, like wearing more sun-protective clothing.

Elias Quijano, a principal at venture capital firm Northpond Ventures, asked whether Smith and Kruse had looked into the mechanism of action for other effects aside from melanin production. He wondered whether there were other activity or repair pathways that the mechanism displayed.

Smith said early research did seem to show that it protected skin from cancer in ways besides darkening the skin.

Gunes Bozkurt, director of venture investments at Beiersdorf, wondered about modulating the dosage of the drug. What if people experience more pigmentation than they desired or uneven skin tone changes?

Smith said that different cases had different priorities. Those suffering from a debilitating illness would likely not care as much about cosmetic imperfections — provided the treatment was effective.

In the distant future, if the drug was approved for cosmetic purposes, he thought people would likely want a version that had a shorter half-life so they could regularly adjust the dosage and their appearance.

Chaya Patel, a principal at Mission BioCapital, asked whether a future trial would compare their drug to the implant or a placebo. Smith said he’d love to run the drug head-to-head against a competitor — an option he thinks would be possible considering the similar primary endpoints his drug and existing ones would target in relatively small trials.

As for questions about what indication the drug would target first, Smith cited XP as a possibility. While the disease is extremely rare in the U.S. (with about a 1 in 1 million incidence), it is more common in Japan (about 1 in 20,000), where patients are funneled to dedicated centers upon diagnosis.

Smith imagined a “very small, focused trial” in Japan — helping the drug get approved in a specific indication before expanding to other uses, including potential cosmetic ones.

Unlike on “Shark Tank,” there was no dramatic conclusion. Smith didn’t have to sell his company, partly due to the fact that no company yet exists. But much of the feedback and suggestions that emerged from “Guppy Tank” will likely prove valuable as he and Kruse move along the path of drug development.

For Ed Doherty and his collaborators, the vaccine project has been a long haul.

“It’s going to kill me, but it’s the best thing I’ve ever done,” joked Doherty, co-founder of Attivare Therapeutics, an anticancer startup born out of Harvard labs amidst the pandemic’s disruptive swirl. 

Attivare’s technology essentially creates tiny “factories” in the body where immune cells learn to recognize tumor cells and attack them. The process has yielded promising results in cancer clinical trials and been extended to fight infectious diseases, including a COVID vaccine created while the team was still at Harvard but too late in the pandemic to be part of the global response.

And in December of 2025, the Natick, Massachusetts, company received a $6.6 million grant extension from the Gates Foundation to continue to develop a vaccine for malaria, one of the world’s deadliest infectious diseases.

The story of the evolution of the vaccine project illustrates the potential power of different teams — basic scientists, translational scientists, and clinical medical researchers — working collaboratively. It also highlights how the path to discovery is seldom a straight line.

In fact, the story of WDVAX, as the cancer vaccine was initially called, doesn’t begin at the lab bench and move directly to patient bedside. Instead, the project wandered from bench to bedside, then back to the bench. There, additional improvements led Doherty to leave the Wyss Institute and establish Attivare, which is seeking once again to get the vaccine to patients.

And Doherty wasn’t working alone. 

Attivare’s tiny “factories” are made of a porous biomaterial via a process pioneered in the lab of David Mooney, Robert P. Pinkas Family Professor of Bioengineering in the John A. Paulson School of Engineering and Applied Sciences. It was further developed in Mooney’s lab on Harvard’s Longwood campus, at the translation-minded Wyss Institute for Biologically Inspired Engineering. 

“There’s a two-way street here,” said F. Stephen Hodi, a Harvard Medical School professor of medicine at the Dana-Farber Cancer Institute who ran the clinical teams during the WDVAX trial. “Having these teams and connections going from bench to bedside, back to bench and such, I think will speed up our ability to help patients.”

It’s been a long road for Doherty and colleagues. It’s been perhaps a longer road for the technology itself, which grew out of advances in Mooney’s lab at the University of Michigan in the late 1990s. 

There, he developed a gas foaming process to create porous structures to house or transport proteins. He mixed in small particles of salt and sugar, which dissolve when washed with ordinary water, leaving behind a porous scaffold made of the biodegradable material used in medical sutures, which releases its contents as it slowly degrades.

At that early stage in the technology’s development, the idea was to use the device in tissue engineering, during which scientists alter cells to correct genetic or other defects and then implant them in the patient to replace the patient’s own defective cells. 

Mooney envisioned using the structure to slowly release proteins known as growth factors at the transplantation site, which would encourage the growth of transplanted cells. 

The idea of a cancer vaccine came later, after Mooney moved to Harvard in the mid-2000s, in discussions with Ph.D. student Omar Ali, who had followed Mooney to his new post. 

“We developed this as a way of gently encapsulating proteins in this polymer,” Mooney said. “When we started thinking about vaccines 10 to 15 years later, we realized this would be a nice technique to use.” 

In the early 2010s, Mooney became a core faculty member at the Wyss Institute for Biologically Inspired Engineering. He discussed the idea with Glenn Dranoff, a cancer vaccine expert and then a Wyss associate faculty member. 

“We started talking with Glenn, who was then co-leader of Dana-Farber Cancer Institute’s Cancer Vaccine Center. He had been developing cancer vaccines and putting them in clinical trials for probably a couple of decades by that point in time,” Mooney said. “He basically said, ‘We should do a clinical trial.’ And I said, ‘We should do a clinical trial?’ And he said, ‘Yeah, we should do a clinical trial.’”

The researchers launched into mouse studies that would ultimately total nearly 100. Though they would take the next seven to eight years to complete, they’d also confirm Mooney and Glenn’s hunch that the technology had the potential to be a potent tool in oncologists’ anticancer arsenal.

“We were able to generate really robust immune responses in a variety of models of different types of solid tumors,” Mooney said. “We noticed that the strategy was able to not just halt cancer progression in these mouse models but also cause complete regression for a sizeable fraction of the animals. That was a very encouraging result and suggested that this was a strategy that really robustly altered the immune response against these cancers.” 

But before the human trial of WDVAX could kick off in 2013, there was work to do by the Wyss Institute’s translational scientists, led by Doherty. He had worked with biomaterials in industry for 25 years, and saw his task as standardizing the technology and making it replicable, a requirement for any medical product.

He also loaned his expertise in regulatory affairs to guide the clinical trial application through the Food and Drug Administration.

Doherty and his team of 18 experts in quality control, manufacturing, and analytic chemistry worked with Dranoff and Hodi at Dana Farber, which led the clinical side of the trial, and Jerome Ritz, HMS professor at Dana-Farber’s Connell and O’Reilly Families Cell Manipulation Core Facility, which created the biological materials needed.

Doherty’s team fine-tuned the device and its immunologic approach. They built equipment that could be transferred for the trial to the CMCF, which manufactures genetically engineered cells for patients enrolled in early phase clinical trials. The CMCF adapted the process used in mice into one that could be used in humans and then in manufacturing WDVAX. 

“This is all about reproducibility and safety,” Doherty said. “In order to go into humans, they’re going to demand that you demonstrate that you know the pitfalls, you know all the problems, so that when you put it into patients you suddenly don’t say, ‘We didn’t think that would happen.’”

The WDVAX trial, the first to test a personalized, biomaterial-based vaccine, was a Phase 1 trial of 21 patients with advanced melanoma that had spread throughout the body. 

Since the main purpose of a Phase 1 trial is to evaluate safety, its patients were enrolled one after the other. That slowed progress but ensured any negative effects would be revealed and addressed before it was given to another patient. 

Clinicians biopsied the subjects’ tumors, divided the biopsied tissue in two, then sent half to the lab for clinical analysis and half to technicians in the CMCF to make the vaccine. CMCF technicians put the tumor cells through multiple freeze and thaw cycles, which broke down tumor cell walls and created a mix, called a “lysate,” of tumor proteins which were then incorporated into a WDVAX product for each patient.

The vaccine, inserted surgically under the skin, is a spongy tablet about the size of a baby aspirin. It is seeded with tumor lysate and molecules designed to spur an immune response. One, GM-CSF, attracts dendritic cells, key immune system cells that sound the alarm when they encounter antigens like the tumor proteins in the lysate. Others, CpG oligonucleotides, are adjuvants whose purpose is to activate the dendritic cells. 

Once exposed to the lysate, the activated dendritic cells carry information about the tumor from the vaccine site to the patient’s lymph nodes. There, the cells encounter immune system T-cells, which the dendritic cells activate, sending them out to attack the tumors.

The study was detailed in the journal Cancer Immunology Research last July. Hodi and others involved in the trial said it went well, particularly with regard to its primary measure of safety and in researchers’ ability to consistently fabricate the vaccine.

In addition, the vaccine elicited an immune response, albeit one not as robust as in mouse trials, and 43 percent of patients exhibited stable disease.

But researchers were dissatisfied with the time it took to manufacture the vaccines, which had to be made individually for each patient, tailored to their specific tumor, and then tested. Taking weeks to create a vaccine was too slow for those with a fast-moving disease. 

And, though the vaccine was determined to be safe, six deaths due to cancer progression were observed.

“After extensive training and developing detailed manufacturing procedures, we have to prove to ourselves and the FDA that we can do this successfully and consistently,” Ritz said. “So, we require a minimum of three ‘validation’ runs to manufacture vaccines that meet all release criteria but don’t go into the patients.” 

The results left researchers thinking about ways to improve the vaccine to achieve the more dramatic results seen in mice.

One observation was that “checkpoint proteins,” which slow T cells’ attack on cancer cells, were upregulated, essentially being produced in greater quantities. Perhaps, they said, the vaccine could be deployed in conjunction with checkpoint blockade therapy, which inhibits checkpoint proteins, and might free the T cell attack. 

Another improvement was already being developed. An injectable version, which would be much easier to administer, was developed in Mooney’s lab in 2015. The injectable vaccine uses tiny silica rods instead of the foam-based bioscaffold. The rods, once inside the body, self-assemble into a bioscaffold that provides the same functional structures as the pores in WDVAX. 

“If you think about doing this routinely, at clinics all over the place, we realized that we needed to make this injectable, consistent with how we get shots today,” Mooney said. “You get a shot, put a Band-Aid over it, and walk out the door.”

Enthusiasm about the technology’s growing promise prompted Doherty to start Attivare in 2021 with others from Wyss, including the former Director of Entrepreneurship Jessica McDonough, senior staff scientist Fernanda Langellotto, and scientist Ben Seiler. 

Despite their enthusiasm, however, pandemic-era shutdowns created hurdles during the company’s launch. 

“During the shutdown, I had left the Wyss and spent all of my time starting the company in my basement,” Doherty said. 

Doherty and other co-founders endured, and Attivare eventually received $7.7 million in seed funding. 

In the years since, they retuned the vaccine toward infectious diseases, with the Gates Foundation’s interest sparked by the vaccine’s ability to slowly release immune stimulating molecules over time, extending and boosting immune response. 

The company is also working on an AI model to predict how individual patients’ cancers will respond, which will boost chances for therapeutic success. 

Supporting it all, the funding environment finally seems to be improving, Doherty said. He hopes Attivare will attract $20 million to $30 million to support the AI-Oncology Program this year. 

As he and colleagues at Attivare work toward a vaccine that will help patients, Doherty keeps in mind the practical vision of Wyss Institute founder Hansjorg Wyss, who believes promising technology shouldn’t languish in a lab. 

“Mr. Wyss was a big advocate of the idea that really great technology that doesn’t get to the patient isn’t really great technology,” Doherty said.

Deep into Paolo Pasco’s audition process for “Jeopardy!,” his mother revealed to him a bit of “mom lore.” Decades earlier, she, too, had auditioned for the quiz show.

Back then, all auditioning took place in person. In his 2006 book “Brainiac,” Ken Jennings recalls driving 12 hours from Salt Lake City to Sony’s Los Angeles studios just to take a preliminary test with a friend. Later, they repeated the drive for screen tests. (Jennings, now the show’s host, went on to win 74 straight games and $2,520,000, easily covering gas money.)

Much has changed. Pasco, a 2022 College graduate, is fairly sure he took his assessment in a mall food court, alongside three friends, on separate phones, who were also taking the test. “We might as well do it now,” he thought to himself.

Pasco later won seven straight games before dominating the show’s Tournament of Champions. In August, he’ll return with other members of the uber elite for “Jeopardy! Masters.”

Though his work as a crossword creator helped build his knowledge base, Pasco was struck by the expansive world of online quizzing he discovered as he prepared for the show: competitions like Quizzing in Progress, LearnedLeague, and Online Quiz League; communities that offer tips and organize practice games; and troves of flashcards that contain nearly every “Jeopardy!” question ever asked.

Part of his preparation resembled that of past contestants. Knowing his weakness at identifying bodies of water, he drilled the subject until he considered himself passable. Online resources helped. His girlfriend even made a website of curated quizzes, the Triviary, to hone his aquatic identification skills. (Pasco notes that other contestants who have appeared on the show have also made use of the site.)

As a lifetime puzzle fan, he’s seen the same growth in crosswords. “That community has really blown up,” he said, “especially in a post-Wordle world.”

“At least for me, the best way to learn something is getting it catastrophically wrong somewhere else.”

Enrollment numbers for this year’s American Crossword Puzzle Tournament were so large that organizers cut off registration two months before the event. Next year, the competition will move from its longtime home in Stamford, Connecticut, to accommodate more competitors.

Pasco’s crossword timing couldn’t have been better. As Wordle surged to mainstream popularity, Pasco was grinding out puzzles for the biggest crossword publishers. “Suddenly, all of the outlets saw the potential in short and casual daily-style games,” he said. He’s ridden the wave to postgraduate jobs at The Atlantic, LinkedIn, and TED Conferences, where he now serves as games editor.

Among trivia competitors, he said, he’s still in the “dewy-eyed, new-to-the-world phase.” A few weeks ago, he dragged himself in front of the computer in the early morning to answer 240 rapid-fire questions for the World Quizzing Championships. Though he improved considerably from two years ago (from 71 correct answers to 118), he expressed awe at the range and depth of the knowledge required — and admiration for seasoned quizzers such as fellow “Jeopardy!” champion Victoria Groce, who scored 191.

“The back half of the quiz was things that were completely new to me,” he said. “They had never come across my desk before.”

(If you can identify the vulture named after a 19th-century German naturalist considered to be the world’s highest-flying bird, consider competing next year.)

Reaching that level of mastery may require plunging further into the trivia depths, with failure as a prod: “At least for me, the best way to learn something is getting it catastrophically wrong somewhere else.”

As he prepares for the markedly more difficult questions asked on“Jeopardy! Masters,” he takes solace in the idea that the only solution “is just to learn everything.” He knows that there are levels to the obsession. However invested he is, he said, “There are people who are five times more in the paint than you.”

And if his latest “Jeopardy!”endeavor falls short, he’ll always have crosswords.

Why do some urban spaces have character while others feel soulless? 

For the latest installment of “One Word Answer,” Stephen F. Gray, associate professor of urban design at the Harvard Graduate School of Design, parses the qualities of an “authentic” neighborhood. 

---

Can you make brand-new authentic? It’s an interesting question, and the answer turns on a few things at once: how a place is built, who builds it, who gets to stay, and whether the experiences inside it connect to one another or just sit side by side.

Start with how it’s built. An authentic city is heterogeneous. But take the Seaport neighborhood of Boston, for example. A group sat down and contrived it, as opposed to an authentic neighborhood where there are so many experiences mixed up together, and each of them true to themselves — and more interesting in relation to each other. There’s turnover, there’s a fire, a family dispute, all these things happening that result in what we see today. What we think is an appropriate part of that mix is a question of values, though.

Heterogeneity is what developers reach for when they’re trying to manufacture authenticity, but they tend to reach for the surface rather than the substance. Developers, when they’re trying to pitch something new, want it to seem authentic to the place. For example, there’s a large area in Miami — Wynwood — that was redeveloped. It was all these warehouses, all this graffiti, and the developers essentially took that as the theme. So all these new condos and developments, which are very vertical and glassy and steely, are also very referential to that graffiti street culture, but only superficially. Now, that’s an attempt at authenticity. But who gets to benefit from that? Is it the graffiti artist who barely makes rent? Usually the answer is no. And in that case, is that authenticity, or is that appropriation? 

Pull back from any one project and a pattern emerges. There’s territorial inequality across cities. Usually the places where capital accumulates are the places where the people who make a city authentic can’t afford to be and don’t see themselves reflected; and the places where they do see themselves reflected are the places that are less capitalized. The new district gets built; the neighborhoods that already have what that district is trying to simulate go without investment.

The big question for developers, planners, and the people who live in these neighborhoods is, how do you bring that kind of authenticity and welcoming and belonging to the vast communities that make a city authentic and interesting and fun? And how do you bring opportunities and resources into areas that already have all of that great authenticity, but lack investment? 

“Authenticity describes both the buildings and the people inside them, which means giving the people who live there, or will live there, the standing to shape what gets built.”

The honest answer is that you can’t theme your way into it. Authenticity describes both the buildings and the people inside them, which means giving the people who live there, or will live there, the standing to shape what gets built. In my practice, when we’re engaging with communities about neighborhood change, it’s not a one-way street; it’s a dialogue. They’re giving us information about what needs to be addressed and what’s happened in the past and how they see the future. But we need to give them information too, about how these things actually come to pass. 

For instance, someone may say they want more affordable housing. So we say, OK, because of Boston’s inclusionary development policy, to make a project work you need 13 percent of units to be affordable and the other 87 percent to be unaffordable. If you want 100 new affordable units, you’ll also get nearly 800 units at market rate.

So the difference between urban spaces that have character, vitality, and life to them, and others that don’t? An authentic city is more than the sum of its parts. It’s not just a gimmick where you can take a selfie to show your friends that you were there. There’s a sort of connective and episodic movement through a set of experiences. There’s the idea of time and incrementality that results in a heterogeneity of not only individual buildings and spaces that are built and changed and reimagined and covered over and combined, but also of the people who make, inhabit, and activate the city. And there’s the agency that lets those people shape the places they live in.