The aim of such policies is straightforward but ambitious: to keep resource use within planetary boundaries while maintaining a high quality of life. Although much work on sufficiency has focused on cities—with measures such as mixed-use neighbourhoods or urban agriculture—rural areas have received far less attention. To fill this gap, the researchers visited 46 Swiss municipalities and surveyed local councillors about their sustainability goals, initiatives, and the barriers they face.

Creative solutions

Their findings reveal an array of creative approaches, particularly in mobility. In some villages, for instance, free bus services link neighbouring communities, while tourists ride public transport at no charge—funded by slightly higher visitor taxes. In other places, residents’ proposals, such as community-driven climate strategies, have spurred broader sufficiency efforts.

Political mandates can also catalyse change. Municipalities that write resource efficiency into local regulations, or that join national programmes like Energy City, create a framework for sustainable decision-making. Pilot projects, meanwhile, help make the benefits tangible. A trial 30 km/h zone, initially seen as intrusive, may later prove popular if it improves safety and livability.

Effective communication and participation are vital. When officials discuss a proposed cycle path during a public walk, for example, they both build trust and ensure the project reflects local needs. Yet limited capacity remains a challenge: small municipalities often lack staff, time, and funds. Federal or cantonal financial support can help, as can the commitment of motivated individuals.

Broad aims

Interestingly, many of the 500 measures documented were not explicitly labelled as “sufficiency” policies at all. Instead, they pursued other aims—such as strengthening local economies by sourcing food for schools and care homes from nearby farms—which incidentally reduced transport distances and resource use. Only two municipalities in the sample used the term “sufficiency,” yet many were already practising it in spirit.

The researchers distinguish between three approaches to sustainability. Efficiency means doing more with less—reducing the resources required to produce or consume a given good. Consistency focuses on closing resource loops and designing products and services to be environmentally benign. Sufficiency, by contrast, entails consuming and producing less altogether, relying on slower, more local systems: cycling instead of driving, eating regional food, repairing rather than replacing.

Sufficiency policy, then, seeks to create the conditions under which such choices become natural and rewarding. Grounded in the concept of planetary boundaries—the thresholds beyond which humanity endangers Earth’s stability—it offers a path for rural communities to prosper without overburdening the planet.

“Innovators encounter setbacks all the time,” note the authors. “Challenging the status quo is difficult work.” Their study followed hospital-system employees in the northeastern United States who were trying to improve internal processes. When their efforts stalled—prompting frustration and self-doubt—they often turned to colleagues for what the researchers term “interpersonal holding”: emotional support that helps people contain distress and make sense of adversity.

Kinds of support

The study identifies three broad forms of such support. Tight holding comes from close colleagues who offer empathy and help to ease stress. Loose holding is provided by more distant peers who nonetheless encourage persistence. Denied holding, by contrast, occurs when others dismiss or amplify the innovator’s frustration, hastening abandonment of the effort.

Support, the researchers found, is not confined to any particular rung of the hierarchy. Managers, peers, or subordinates may all act as “holders”—or fail to do so. The key lesson is that innovation is arduous, often lonely work. Organizations that expect staff to disrupt established routines must ensure those employees are not left to do so unsupported, and that they understand what genuine, useful support looks like.

Research from Northeastern University suggests that the success of populist parties, such as the PVV, stems less from their own brilliance than from the normalization of their ideas by mainstream politicians. By adopting tougher rhetoric on immigration and security, centrist parties have inadvertently helped make the far right appear respectable.

Made mainstream

For more than a century, the Netherlands’ fractious, multi-party system has relied on coalition governments. Mr Wilders, a peroxide-haired provocateur known for his anti-Islam tirades, was part of the most recent four-party coalition—an experiment that lasted a mere 11 months. He withdrew his support after his partners rejected his maximalist immigration demands, including a total ban on asylum applications and the deportation of Syrian refugees. That collapse forced a new election barely two years after the last.

Mainstream parties are now attempting to contain Mr Wilders by reviving a cordon sanitaire—a refusal to govern with extremists. The centre-right People’s Party for Freedom and Democracy (VVD), which held the premiership from 2010 to 2023 and briefly shared power with the PVV, has vowed not to work with him again. Yet such political quarantines are fraying across Europe and are only “minimally holding” in the Netherlands. The previous coalition marked the first time the PVV had been invited into government.

Unless the PVV converts its polling lead into an outright majority—a rarity in Dutch politics—the mainstream parties will again be forced to cobble together a coalition and resist the temptation to rely on Mr Wilders for support. That will test their unity as much as their principles.

Playing into the right’s hands

The researchers warn that the fixation on immigration and security plays into the far right’s hands. To blunt populist momentum, they suggest, mainstream parties must address more pressing concerns: inequality, climate change, the cost of living and creaking infrastructure.

Two parties are currently vying for second place: a rejuvenated Christian Democratic Appeal (CDA) and a joint GreenLeft–Labour alliance. Both have made housing central to their campaigns. The CDA’s leader, Henri Bontenbal, promises a return to “normal, civilised politics”. Yet even housing—a long-standing Dutch headache—has been seized upon by the PVV, which blames migrants for the shortage. The issue may well become the next electoral battleground.

The rise of the far right, the study notes, reflects a deeper malaise. Many Dutch voters feel frustrated and fearful, symptoms of a wider Western discontent. The Netherlands, often portrayed abroad as a bastion of liberal tolerance, may be less so than imagined.

Uncertain future

Whatever the outcome, the country faces months of horse-trading. Coalition negotiations after the 2023 vote dragged on for eight months before a government was finally installed—the most right-leaning in post-war history. Long admired for producing consensus, the Dutch system now looks increasingly ponderous as the economy grapples with sluggish growth and the challenges of the technology and AI revolutions.

While populists elsewhere—most notably Donald Trump—promise swift, decisive action, the Netherlands remains bound by its tradition of careful compromise. Whether that restraint proves a strength or a weakness will soon be tested again at the ballot box.

To manage this flow of ideas, many corporations have created dedicated “scouting” units. These are staffed with professionals who straddle the line between external innovation and internal bureaucracy, charged with spotting promising technologies and shepherding them through the organisation’s product-development maze. Yet new research from INSEAD suggests that this supposedly seamless bridge between start-ups and corporates is far trickier to maintain than it looks.

Creating the bridge

The study examined knowledge scouts at a large Silicon Valley technology firm, whose multi-divisional structure reflects that of many global corporations. Scouts were not tied to particular divisions but were expected to support several simultaneously. Their success was measured not by the number of start-ups they identified, but by whether their finds led to new products.

The findings reveal a paradox. The more cross-divisional experience a scout gained, the less likely they were to succeed in launching new products within any given division. Deep familiarity with one division’s people and processes proved useful locally but did not travel well. When divisions had prior exposure to similar technologies, scouts’ experience transferred more easily; when they competed directly with one another, that experience often backfired.

Interviews with scouts and managers revealed why. Knowledge scouts spoke of steep learning curves and political minefields when moving between divisions. Relationships, trust, and tacit knowledge—so vital to getting things done—rarely survived the leap. As one scout put it, “The person who’s your go-to contact in one division might be the one you should avoid in another.” Copy-and-paste strategies seldom worked.

Their task became simpler when the receiving division was already familiar with the type of external technology in question. “The first cases typically faced a lot of resistance,” said one manager. “But once the novelty was accepted, it created an entry point for others to follow.” Rivalry between divisions, by contrast, made collaboration harder. Scouts with experience in competing areas often found their involvement unwelcome, as colleagues feared losing ownership of ideas or budgets.

Specialization vs versatility

The study highlights a dilemma at the heart of corporate innovation: the tension between specialisation and versatility. Deep expertise within a division builds credibility but limits mobility; broad experience across divisions spreads knowledge but dilutes influence. This trade-off undermines the very premise on which scouting units are built—that a handful of well-connected intermediaries can ferry fresh ideas smoothly across sprawling firms.

The authors suggest several remedies. Firms should expose scouts to multiple divisions, but ideally to those that do not compete directly, and where similar technologies are already in play. They might also broaden the remit of scouting units, assigning them projects that span several divisions to encourage collaboration. Since scouts often lack formal authority, companies could invest in training to hone their influence, negotiation, and political skills.

Scouts, the study concludes, are not mere couriers ferrying ideas from start-ups to boardrooms. They are diplomats, translator,s and brokers of innovation—roles requiring not just technical nous but political dexterity. If firms can balance depth with breadth in these positions, they may yet unlock the full promise of the start-up–corporate marriage that so many have been chasing.

The study, which was undertaken across the US, Germany, France, Australia, Japan, China, and the UK, found that around a third of organizations were looking to AI before they hired anyone, with many expecting this trend to grow in the years to come.

Maintaining confidence

Despite this, the workforce remains pretty optimistic. At least that’s the picture painted by research from the University of California, Merced, which suggests that the public isn’t unduly concerned about AI affecting their livelihood.

The study found that people were generally optimistic about their chances when faced with AI entering the workplace, with this optimism especially strong in the longer term.

The researchers surveyed a few thousand U.S. adults, each of whom was given information about the rapid developments in generative AI. After reading about the technology, it was common that people thought the tech would arrive sooner rather than later.

Despite that, few thought that either the economic outcome would worsen or that policymakers should offer new forms of support to people. In other words, they weren’t really any more concerned about technological unemployment.

Stubborn optimism

“These results suggest that Americans’ beliefs about automation risks are stubborn,” the authors explain. “Even when told that human-level AI could arrive within just a few years, people don’t dramatically revise their expectations or demand new policies.”

This should perhaps come as no great surprise. After all, a recent Spanish study showcased what the researchers refer to as the “invulnerability bias,” which basically suggests that we tend to think that automation is something that will affect other people more than it will ourselves.

The study builds on a 2023 survey from Pew Research, which found that 62% of U.S. adults believe AI will have a major impact on workers generally, but only 28% think it will significantly impact their own position.

Interestingly, however, the Spanish researchers found that while people were confident that their job would survive, they weren’t confident that it would be as enjoyable or meaningful.

Taking the threat seriously

While the precise impact of AI remains uncertain, it seems better that people are aware of the threat rather than unduly optimistic. Here, the two studies provide a degree of insight into what might move the needle. While the University of California tried to raise awareness of AI to a slight extent, the Spanish study was more direct in identifying the correlation between AI literacy and greater awareness of the potential risks of automation.

It’s likely that this awareness needs to be practical and experiential rather than theoretical. What’s clear from the Californian study is that making technological threats feel more immediate isn’t enough to either provoke action individually or mobilize support for greater government action.

This creates an uncomfortable paradox for policymakers and business leaders. If public concern only materializes after disruption becomes personal and widespread, societies may find themselves scrambling to build safety nets and retraining programs precisely when they’re most urgently needed, and when budgets are most strained by rising unemployment.

Increasing exposure

The research suggests that neither alarming predictions nor distant timelines will prompt preemptive action. Instead, the pathway to realistic assessment may lie in direct engagement: getting workers to actually use AI tools in their domains, to see both the capabilities and limitations firsthand, and to develop informed rather than abstract opinions about their own vulnerability.

Whether this stubborn optimism proves justified—whether new jobs truly will emerge as old ones disappear, as has happened in previous technological transitions—remains to be seen. What’s certain is that we’re conducting a live experiment in how societies adapt when the workforce remains calm in the face of a transformation that experts insist is imminent.

For now, most workers are placing a quiet bet: that they’ll figure it out when they need to, that adaptation will come naturally, that this time won’t be fundamentally different. History will judge whether that confidence was wisdom or wishful thinking.

The semester-long study compared two sections of an upper-level antitrust economics course, taught by the same professor with identical lectures, assignments, and exams. One group of 29 students was allowed—indeed encouraged—to use tools such as ChatGPT under structured guidance and with mandatory disclosure. A parallel section of 28 students was barred from AI use but received similar support through traditional study aids.

Minimal impact

The outcome? No discernible difference in exam scores or final grades. Yet the AI-enabled students were consistently happier and more involved. They participated more in classroom discussions, focused their study into longer but fewer sessions, and showed signs of more reflective learning—editing AI outputs, spotting errors, and often preferring their own answers over the machine’s.

“AI didn’t make students smarter,” the researchers conclude. “It made them more efficient and confident.” Those with AI access spent less time preparing for exams and homework yet felt more in control of the material.

Standard course evaluations echoed the sentiment. Students in the AI section rated the instructor’s preparation and use of class time notably higher than their non-AI peers. They were also far likelier to express interest in pursuing careers involving artificial intelligence.

Some reassurance

For anxious academics, the findings offer a measure of reassurance: integrating AI need not undermine academic standards. The authors advocate a “permit with scaffolding” approach—allowing AI use within clear boundaries and with explicit instruction on when and how to deploy it. “Letting students engage with it just creates a different environment,” they argue.

To avoid bias, the researchers assigned the AI-permitted condition to the afternoon section, which had historically performed slightly worse—a conservative test of the technology’s impact. All students took paper-based exams without access to notes or devices, ensuring any performance gains were genuine.

Still, the authors caution against sweeping conclusions. The experiment involved only 57 students, and many of the most positive outcomes—such as confidence and engagement—were self-reported. Yet the results hint at a future where AI becomes less a threat to learning than a partner in it. As in so many other professions, the question for educators may no longer be whether to use AI, but how best to manage it.

Research from Brown University shows how dangerous this is and how roughshod most chatbots are when it comes to basic ethical standards. The study compared how chatbots operate with standards from bodies such as the American Psychological Association, and found the bots to be severely lacking.

Breaking the rules

The study found that chatbots respond in a number of ways that aren’t recommended by official bodies. For instance, they might provide responses that reinforce the negative beliefs users have about themselves, or create a false sense of empathy with users.

“In this work, we present a practitioner-informed framework of 15 ethical risks to demonstrate how LLM counselors violate ethical standards in mental health practice by mapping the model’s behavior to specific ethical violations,” the researchers explain.

“We call on future work to create ethical, educational, and legal standards for LLM counselors—standards that are reflective of the quality and rigor of care required for human-facilitated psychotherapy.”

Mental health support

Research has shown that around half of users with a mental health condition use chatbots for support. The Brown researchers wanted to test how helpful that was. They input a wide range of different prompts to explore the kind of responses chatbots typically provide. Rather than aiming to trip up the bots, the researchers accept that users are turning to AI for support and want to see how the bots can do better. The aim was to see if different prompts provide better answers.

“Prompts are instructions that are given to the model to guide its behavior for achieving a specific task,” the researchers explain. “You don’t change the underlying model or provide new data, but the prompt helps guide the model’s output based on its pre-existing knowledge and learned patterns.”

This is obviously slightly easier to achieve than changing either the data or the model. It requires users to actively frame the conversation in a professional way, such as asking the bot to behave as a cognitive behavioral therapist might when reframing our thoughts.

“While these models do not actually perform these therapeutic techniques like a human would, they rather use their learned patterns to generate responses that align with the concepts of CBT or DBT based on the input prompt provided,” the researchers continue.

Smarter responses

The researchers note that individual users are already experimenting with prompts to guide chatbots toward more therapeutic responses. Many share their preferred phrasing on social media platforms such as TikTok, Instagram, and Reddit, where communities exchange advice on how to elicit more empathetic or helpful replies. But the implications of prompting go far beyond casual experimentation.

A growing number of mental health apps are built on top of general-purpose large language models that have simply been “steered” with mental health–related prompts. Understanding how these prompts shape the behavior of such systems is therefore essential if they are to be used safely and effectively.

To explore this, the researchers observed a group of peer counselors working with an online mental health support platform. Seven counselors, all trained in cognitive behavioral therapy techniques, conducted self-counseling sessions with LLMs prompted to emulate CBT practitioners.

Not meeting the grade

A smaller subset of sessions was then assessed by a group of trained clinical psychologists. They tried to compare the exchanges people had with AI against the standards and ethics they adhere to in their professional practice. In total, this analysis revealed 15 ethical risks that fell into five broad categories:

• Lack of context, with the AI often applying a one-size-fits-all approach and not taking into account the users’ particular circumstances.
• A one-way dialog, with the AI seldom engaging in a more collaborative approach to problem-solving. The AI would also often reinforce harmful self-beliefs that the user had.
• The faux empathy AI exhibits has been well documented, and this was evident in the research, with the chatbot trying to simulate a genuine connection.
• The chatbot also showed clear gender, cultural, and religious biases, which made its responses discriminatory.
• Last, but by no means least, it struggled to respond effectively to crisis situations, such as suicidal ideation.

Suffice it to say, human therapists are by no means perfect either, but the scale of adoption of generative AI for mental health-related conversations makes this an area that needs to be urgently addressed, not least because professional therapists are accountable in a way that the AI companies are, thus far at least, not.

The researchers emphasize that their findings should not be interpreted as an argument against the use of AI in mental health support. Instead, they see the technology’s potential to expand access to care—especially where cost or availability remain barriers. What’s needed, they argue, is thoughtful design, robust oversight, and clear ethical standards to ensure AI systems do more good than harm.

“There is a real opportunity for AI to play a role in combating the mental health crisis that our society is facing, but it’s of the utmost importance that we take the time to really critique and evaluate our systems every step of the way to avoid doing more harm than good,” the researchers conclude. “This work offers a good example of what that can look like.”

This phenomenon is no less pronounced in medicine, as a recent study from Johns Hopkins shows that doctors who use AI at work are also deemed less competent by their peers.

Adoption matters

Often, when we think about generative AI, we think about the capabilities (or lack thereof) of the technology. Vendors talk about having a PhD in your pocket, while naysayers worry about its tendency to hallucinate.

When it comes to adoption, however, there are various very human considerations that are just as important, if not more so, than the technical specifications. For instance, it’s widely known that many employees fear automation, so they will resist the introduction of technology they believe will either remove their job entirely or strip it of meaning.

We’ve seen in various studies that AI can help to improve clinicians’ diagnoses, but the Johns Hopkins study reminds us that this efficiency improvement can carry social costs. The study found that doctors who leaned heavily on AI in their clinical decision-making were viewed by peers as less competent and less capable overall. In other words, their reliance on AI didn’t just raise questions about the technology, it raised doubts about them.

Viewed with skepticism

“We already know AI is part of medicine’s future,” the researchers explain. “What’s striking is that doctors who use it in their decisions are seen by colleagues as less capable. That kind of stigma—not the technology itself—may be what slows progress.”

In a randomized experiment, 276 clinicians were asked to evaluate scenarios where doctors either made decisions independently, used AI as their main decision-making tool, or used it to double-check their judgments. The more a doctor depended on AI, the greater the “competence penalty” they faced in the eyes of their peers.

“In the age of AI, human psychology remains the ultimate variable,” adds Haiyang Yang, first author of the study and director of the Master of Science in Management program at Carey. “How people perceive AI use can matter just as much as how well the AI actually performs.”

The paradox of progress

Interestingly, while doctors who avoided AI altogether were viewed most favorably, most participants still acknowledged that AI can enhance diagnostic precision and improve care quality, particularly when customized to institutional needs.

In other words, it’s very possible that clinicians will value AI in private, but be reluctant to use it or praise it in public so that their capability and credibility isn’t tarnished. As we’ve seen with the other studies, this tension is key to the integration of AI into our daily lives. So long as we run the risk of appearing less competent for using AI, we’re not going to use it, regardless of its actual utility.

It seems inevitable that AI will weave itself into the fabric of medicine, as it has with other disciplines, but its successful adoption won’t be so much down to its technical capabilities as the human factors surrounding it. This was highlighted in a recent paper in Nature, which shows that the key is how humans feel working alongside AI.

Pushed underground

While there remain unquestionable doubts about AI’s efficacy, it also seems unquestionable that it will only get better. What remains equally clear from the last few years is that these human factors have made adopting it into various professions challenging. This perhaps helps to explain why so few AI investments thus far have offered a return on investment. As with so many technologies, experimentation will be key to understanding how the technology can be introduced into workflows effectively. It’s also inevitable that many workflows need to be re-designed to better capitalize on what the technology can do.

None of these things happen if people are scorned for experimenting. All that does is drive the technology underground, where organizations can’t ensure that legal and ethical considerations are taken into account, and where lessons from people’s experiences can’t be learned, as we discourage open conversations.

Ultimately, the future of AI in medicine will hinge on more than data or design. It will rest on fostering a new kind of partnership between humans and machines. One where clinical expertise is not replaced but reimagined, and where progress depends as much on empathy, communication, and trust as on code.

The findings highlight the threadbare nature of Britain’s social safety net and raise questions about the government’s modest plans for an Unemployment Insurance scheme, first outlined in March 2025. The study shows that the UK’s limited unemployment support—both in duration and generosity—amplifies the financial shock of redundancy, while weaker re-employment outcomes deepen the long-term damage.

Sharp drop

In the first year after losing work, British families lose nearly a fifth of their income, compared with 5–6% in the Nordic and continental models. Over five years, the cumulative loss rises to around 35%—more than twice the long-term penalty faced by households in Finland or Germany. “While job loss is a normal feature of modern economies, the financial fallout varies dramatically across countries,” the researchers note. “In the UK, the lack of adequate unemployment insurance leaves families bearing the burden alone.”

Even the industrious efforts of British households fail to close the gap. The study finds that by taking on extra hours or second jobs, UK families offset roughly 22% of their initial income loss. In Denmark and Finland, where benefits are more generous, the equivalent figure is a mere 1–3%; in Germany, 11%.

The shortcomings stem from two structural weaknesses. First, Britain’s unemployment benefits replace barely 20% of previous earnings for up to six months, compared with 60–80% for as long as two years in much of northern Europe. Second, while most Britons do find new work within a year, they typically do so in lower-paid roles—earning around 40% less than before, nearly double the decline seen elsewhere.

Falling short

The government’s proposed Unemployment Insurance scheme is a step forward, but falls well short of European standards. Payments would amount to just a third of the earnings of a full-time minimum-wage worker, far below the 60–80% replacement rates common on the continent.

A more robust system, the study argues, could support a more dynamic labour market, encouraging workers to move between roles without fear of destitution and helping to boost productivity and growth. While greater generosity would carry a fiscal cost, it may prove a sound investment in the long run—especially for a government that has made growth its defining priority.

Such norms vary widely across cultures and have shifted markedly over time. What once counted as proper conduct in one era or place may appear quaint—or even rude—in another. To chart these changes, researchers at the Institute for Future Studies in Stockholm and several other Swedish institutions recently conducted one of the largest cross-cultural studies of its kind, examining how social norms have evolved and how they differ across 90 societies.

Unwritten rules

“Every social situation comes with a set of unwritten rules about what counts as appropriate or inappropriate behaviour,” the researchers note. Some societies are broadly permissive; others, more constrained.

The team surveyed 25,422 participants around the world, asking them to judge the acceptability of 150 everyday behaviours in various contexts. Their results were then compared both across countries and against a similar dataset collected two decades earlier.

The findings point to a simple conclusion: norms are shaped by moral priorities. Societies that emphasise care and liberty—values centred on protecting individuals and their rights—tend to be more permissive overall. Those that prize purity and group cohesion, by contrast, impose tighter constraints, especially on conduct seen as indecent or taboo.

“Societies with more individualistic morality exhibit greater tolerance overall,” the authors explain. “They allow more freedom, particularly regarding behaviour deemed vulgar, but are less forgiving of actions that cause harm or distress.”

Loosening norms

Across the globe, norms appear to have loosened over the past twenty years. People today are generally more accepting of unconventional behaviour than their predecessors were. Yet in more individualistic societies, tolerance has limits: inconsiderate or disrespectful acts are viewed less favourably than before.

In collectivist or morally stringent cultures, meanwhile, the focus remains on preserving propriety and avoiding impurity. The research thus paints a nuanced picture of a world growing freer in some respects, yet firmer in its condemnation of selfishness and rudeness.

Further studies may explore how these moral undercurrents—care, liberty, purity, and cohesion—continue to shape the ever-shifting landscape of human civility.