A recent study from Cornell explores the impact these glitches have, particularly on the hapless person on the receiving end of them. Interestingly, this adverse reaction isn’t because the technical glitches cause us to question the competence of the person experiencing them. Instead, it’s the sense of so-called “uncanniness” that the glitches invoke.

Real implications

Given the widespread use of video calls, the researchers wanted to explore the impact of technical glitches on the participants. The researchers looked at everything from recruitment calls to parole hearings, and consistently found that technical glitches harmed the person suffering them.

For instance, during recruitment calls, experiencing a technical glitch caused the recruiter to negatively assess the candidate. In parole hearings, things were even worse, as technical issues resulted in a 12% fall in the likelihood of someone being granted parole.

“The best feature of video calling is the fact that you basically feel like you’re together,” the researchers explain. “And so when there’s a glitch, you’re right in that danger zone where it’s almost perfect, but not quite – what has become known as the ‘uncanny valley.’ It triggers this switch in your brain where things feel just a little bit creepy.”

Reliable communication

The researchers conducted a number of experiments. For instance, in one, they tapped into the CANDOR (Conversation: A Naturalistic Dataset of Online Recordings) Corpus database of archived video calls. The database consists of a huge number of so-called “get-to-know-you” conversations that are designed to help us understand human conversations.

After each conversation, the two participants were asked to complete a survey that involved understanding the level of connection they felt they had with their conversation partner. The survey also questioned them on whether there were any technical issues during the call. The researchers themselves analyzed each call that featured a technical glitch to understand the nature of that glitch, and whether the glitch affected both parties or just a single person.

The results show that when conversations were affected by a technical issue, the level of social connection achieved was significantly lower. What’s more, this was true regardless of what kind of glitch occurred, and even if the glitch was affecting a single person or both.

Shattering the illusion

So why did this happen? The researchers argue that when we have a smooth video call, it creates the illusion that we’re talking face-to-face with the other person. When there are technical glitches, this illusion is shattered, which then impacts the bond we form with the other person. It’s an ick that’s stubbornly hard to shift.

This phenomenon obviously has even more significant implications when parole hearings and other legal matters are conducted virtually. The researchers looked at the transcripts of virtual parole hearings in Kentucky that took place between January and April 2021 (ie, during Covid). The researchers created a “glitch dictionary” to chronicle the various forms of technical issues experienced during these calls.

Sadly, they found that around a third of parole hearings experienced some technical issues, and these had a profound impact on the parole decisions. The results show that while attendees in glitch-free calls were granted parole around 60% of the time, this fell to 48% when there were technical issues. While we may suggest correlation not equalling causation, the researchers endeavoured to control for the various characteristics of either offender or their crime, so they believe the technical glitches were a big factor.

Unintended consequences

With digital communication so pervasive in the modern world, the results raise real concerns about just how fair this environment is, especially when inequality can mean sections of society have poorer quality equipment and internet connections to work with.

During the pandemic, we flocked to platforms like Zoom and Teams to keep the wheels turning, but we did so without fully understanding the psychological effects of this new way of working. The uncanny valley effect triggered by technical glitches isn’t just an annoyance; it fundamentally alters how we perceive and connect with others, often in ways that disadvantage the person experiencing the technical difficulty.

With virtual calls now pervasive, whether in courtrooms or interview rooms, medical consultations or therapy, the impact of technical glitches could be significant.

The researchers suggest that organizations conducting high-stakes video calls should have protocols in place. For instance, they might reschedule when significant technical issues occur, or at a minimum, make participants aware that glitches may unconsciously bias their perceptions.

As video communication becomes increasingly embedded in our lives, understanding and mitigating these unintended consequences becomes not just important but essential to ensuring fairness and equity in our digital interactions.

It seems a case of the bleeding obvious, yet when we talk about remote work, and especially return-to-office mandates, it often seems like autonomy is the last thing employees are afforded, with managers making decisions for us. Even when hybrid work is offered, it often seems like an arbitrary number of days is plucked from the air rather than consideration given as to what works well for teams and individuals.

What works

It should perhaps come as little surprise, therefore, that a recent Australian study found that it’s less about whether you go fully remote, hybrid, or on-site, but whether people have the autonomy to choose what works best for them.

The researchers tracked just under 500 workers from across Europe and Australia to try and understand how they like to work. The results show that there is often a considerable mismatch between how people want to work and how their employers make them work. Indeed, around a third of employees suffer from this mismatched expectation, with this understandably having an impact on engagement.

The study reminds us that when we feel like there’s a good balance between our personal and professional lives, we’re not only more satisfied at work but also more energized and effective. This also corresponds with a greater willingness to stay at the organization.

“The study found that when the proportion of time that employees spend working from home is aligned with individual preferences, remote work reduces turnover through improved work-family balance satisfaction,” the researchers explain.

Choosing to stay

Of course, it’s one thing to tell a researcher that you’re looking to leave, and another thing entirely to actually up sticks and do so. So, the researchers took this into account and believe that their results point to actual movement rather than just the intent to move. They believe that the connection between autonomy around how and where we work and our desire to stay (or not) is robust and reliable.

I’ve covered the benefits of remote work for many years now, and the evidence is pretty clear that employees like it. It’s also pretty clear that it doesn’t harm productivity. It doesn’t harm collaboration. If done correctly, it doesn’t even harm team bonding. The Australian study reaffirms all of this, with around 3/4 of employees saying they like being able to manage their work/life balance. Similar numbers love saving time on lengthy commutes. Nearly 70% said they were actually more effective at home because they were free from distractions.

While the motto of the research is undoubtedly that managers shouldn’t try and impose boilerplate policies, the majority of workers in the study did prefer a hybrid approach. Notably, just 7% said they wanted to work full-time in the office.

The path forward

The message from this research is clear. Managers should trust their employees to make decisions about their own work arrangements. Rather than issuing blanket mandates or picking arbitrary office attendance quotas, organizations that want to retain talent and maintain engagement should start by asking a simple question: “What works best for you?”

This doesn’t mean complete chaos or the end of in-person collaboration. Most workers in the study wanted some office time. They just wanted a say in when and how much. The solution isn’t choosing between remote, hybrid, or on-site as a one-size-fits-all policy. It’s recognizing that different roles, different teams, and different individuals have different needs.

The irony is that in trying to maintain control through rigid return-to-office mandates, many managers and organizations are achieving the opposite of what they want. They’re disengaging the very employees they’re trying to bring together. They’re driving away talent in a competitive market. And they’re ignoring clear evidence about what actually makes people productive and committed.

The truly sad thing is that we’ve known about the motivational power of autonomy for decades now. Perhaps it’s time to stop treating it as controversial and start treating it as what it actually is: common sense. The organizations that figure this out will be able to attract, retain, and get the best out of the talent they so crave.

A new study from the University of Michigan argues that America’s administrative patchwork—its jumble of cities, towns, and villages, often only a few streets wide—combined with entrenched economic segregation has produced what the authors call “tax-base fragmentation”. Property wealth, they show, is not merely unequal; it is geographically splintered, with some municipalities sitting atop vast pools of taxable value while adjacent jurisdictions scrape by with very little. The result is a stark divergence in what local governments can afford.

Fragmented society

Sifting through 138m property-tax records nationwide, the researchers mapped where wealth lies and how sharply it diverges within metropolitan areas. The gaps are dramatic. They reflect not only household inequality but also state laws that allow affluent enclaves to incorporate as separate municipalities, neatly sealing in their riches.

Two measures illustrate the extremes. At one end sit “municipal tax havens”: small cities whose property wealth per resident far exceeds that of their surrounding metro. Malibu and Miami Beach are predictable examples; suburban Detroit’s Lake Angelus and Bloomfield Hills, less so. Scattered across the country are even tinier places most Americans have never encountered, yet which boast staggering fiscal resources.

At the other end lie “fiscally impoverished” jurisdictions—municipalities whose tax bases are so meagre that funding even basic services becomes a struggle. In Michigan, these include Detroit, Inkster, Flint, and Saginaw. Their predicament is not unique; hundreds of metros display similar chasms between the haves and have-nots.

Driving inequality

The upshot is an overlooked driver of inequality: the way local boundaries entrench advantage. In some states, well-heeled neighbourhoods can simply hive themselves off, hoarding their property wealth rather than sharing it across the region. The surrounding metros may look prosperous on paper, but the lived reality varies block by block and border by border.

To bring these patterns to life, the research team has built an interactive map charting the fiscal capacity of every municipality in the country. It offers a revealing tour of America’s fragmented tax geography—and a reminder that small borders can have outsized consequences.

Researchers at Cambridge Judge Business School set out to reconcile two competing views. One champions creative types as engines of innovation; the other warns that they can be prima donnas who flout norms, demand resources, and sow discord. Creative workers, after all, tend to explore widely and break rules—traits that may spark invention but fray coordination and inflame conflict.

A nuanced picture

Their findings offer a more nuanced picture. Innovation, measured through patent registrations, follows an inverted U-shaped curve: adding creatives boosts output only up to a point. Beyond that, too many idea-generators jostling for attention and resources undermine the team’s ability to execute.

The study also examines how human resource practices shape this curve. Participative management, long touted for empowering staff, can blur job boundaries and complicate coordination. When everyone feels entitled to push their pet idea, territorial scuffles grow. Limited participative management, by contrast, preserves clearer roles and smoother workflows.

Pay dispersion, another popular lever for incentivising performance, fares little better. Though it can stoke individual ambition, it also encourages employees to chase personal glory rather than collective success—a pattern well documented in professional sport, where highly skewed salaries lead to more showboating and less teamwork.

The right mix

The researchers draw on data from South Korea in the early 2000s, when giants such as Samsung and LG were engaged in a talent war that upended the country’s seniority-based labour norms. Using information from the Human Capital Corporate Panel, they classified workers into four categories—creators, trainers, independent performers, and dependent performers—and matched team composition with patent outcomes across 120 single-team R&D departments.

The results point to a simple conclusion: balance, not bloat, is what makes creative teams tick. Teams thrive when a cadre of idea-generators is complemented by those who coordinate, support, and implement. Piling creatives into a room may satisfy managerial instinct, but it is as likely to produce friction as breakthroughs.

The authors urge caution for managers tempted by fashionable HR nostrums. More autonomy, more participation, more differentiated pay—none is inherently beneficial. Each can backfire when applied to teams already dense with strong-minded innovators. In creative work, as in chemistry, the right mix matters more than the raw ingredients.

A recent study from Wharton suggests that technologies like AI could also be playing a role. The researchers argue that a big factor in how much people in the tech industry are paid is their usage of new technologies, like AI. Previous research has shown that women adopt AI at significantly lower rates than men, with Google data showing this is particularly so for those in mid to late career stages.

Lack of experience

The authors argue that it’s not just a case of women not using AI at work though, as they also believe that women are less likely to apply for roles that require a high degree of expertise in emerging technology. The picture is better when they’re actually in those roles, but they’re still less likely to retain their post than men. As these jobs tend to be quite lucrative, it goes some way to explaining the gender pay gap in tech.

Hopefully, it goes without saying that this isn’t due to any kind of deficiency in women. Instead, the authors highlight the structural barriers that keep women out of lucrative jobs in the tech industry.

The study suggests that many high-paid tech jobs are in places where firms spend heavily on technology, which can create a cluster of highly-skilled people. This can disadvantage women who are often less able to relocate, especially when living in dual-career households where moving has wider implications. This can result in many women passing up well-paid opportunities in the sector.

Long hours

The researchers also believe that new technologies require us to learn new skills, which often means putting in extra hours to do so. We know from other studies that men often work longer hours, not least because women do far more of the unpaid labor at home. These family responsibilities can also make it difficult to relocate, which can mean new opportunities are often passed up.

“Staying on the frontier of new technologies depends on learning by doing,” the researchers explain. “If you’re not in those jobs, you cannot pick up the latest skills. That is where the pay gaps start to grow.”

The wage premium is significant, with those working with newer technologies earning around 6% more than those working with older tech.

Making things better

The researchers believe that addressing these structural barriers requires concrete action from employers. Companies should offer childcare support, flexible hours, and options like job sharing to make it easier for people to balance work and family. This would help more women take on jobs that use new technologies.

“More reasonable hours and flexible work schedules would make a real difference,” the researchers explain.

The study added that managers can also help by creating formal training programs that give women the same opportunities as men to learn and work with new technologies. Women often miss out on developing these skills because they are often gained on the job, in roles that are harder for women to take.

If the learning process is formalized, this can level the playing field and ensure that career advancement isn’t dependent on one’s ability to work long hours or relocate at short notice.

Levelling the playing field

By making the acquisition of new skills more formalized, companies can level the playing field and not load the dice in favor of male employees at the expense of their female peers. This should include creating formal mentorship programs, ensuring that training is accessible to employees regardless of their working patterns, and documenting and codifying knowledge so it’s easier to share.

The research reminds us that closing the gender pay gap isn’t just about getting more women into the tech industry, it’s also about rethinking how we organize work to ensure that everyone has equal access to the kinds of opportunities that lead to better pay and drive career progression.

Without these changes, the gap is likely to persist, or even widen, as new technologies continue to emerge and reshape the sector.

“Employee delight”, an awkward-sounding term for a potent emotional state, describes the surge of feeling that occurs when an experience vastly exceeds expectations. It is the sort of moment that can brighten a career and, occasionally, lift corporate performance. Yet, despite its potential, the concept has attracted scant academic scrutiny.

Delighted employees

Researchers at the Universitat Oberta de Catalunya hope to fill that gap. Their work seeks to define the phenomenon, knit together disparate theoretical threads, and offer an empirical framework for studying it.

Delight, they argue, is not akin to the steady glow of job satisfaction. It is a sharp, surprising jolt—an emotional spike triggered by an event that feels both exceptional and personally meaningful. Picture an analyst whose report shapes a strategic decision. Management not only approves her proposal but publicly credits the quality of her work and entrusts her with overseeing its implementation. The unexpected recognition, tied to a significant professional ambition, turns routine satisfaction into something more electric.

Such moments are fleeting, but their consequences can endure. Positive emotional shocks tend to nudge workers towards more constructive behaviour: greater initiative, stronger collaboration, and a heightened willingness to hunt for meaning in their tasks. Over time, they may reinforce deeper qualities—confidence, motivation, and resilience.

The right environment

Firms, the researchers suggest, would do well to cultivate environments where such jolts are possible. This need not entail extravagant perks or grand displays. Rather, it rests on smaller but more meaningful acts: precise and sincere praise, credible opportunities for growth, and leadership rooted in trust.

The study identifies four kinds of circumstances most likely to trigger delight: achieving goals or solving knotty problems; receiving recognition or useful feedback; feeling valued in social or collaborative settings; and encountering unexpected good fortune at work.

Companies often pride themselves on offering “good conditions”. But adequacy, however welcome, seldom dazzles. The real opportunity lies in creating experiences that feel exceptional—moments employees will remember long after the survey-taker has gone.

The analysis draws on an enormous trove of data, from the footprints of 183m buildings to 125,000 household surveys, to map how urban form influences infrastructure access. It models three paths for future expansion and finds that cities that continue to stretch outward rather than build upward and inward will struggle to keep hundreds of millions connected to pipes and sewers. By 2050, a sprawl-driven trajectory would leave 220m more people without piped water and 190m without sewage systems.

Under strain

Much of the developing world is already feeling the strain. In cities from New Delhi and Cairo to Lagos and Bogotá, sprawled neighbourhoods face thinner coverage and steeper bills. Water costs are, on average, 75% higher in dispersed cities than in compact ones, and access to piped water is halved. Residents on the fringes of cities enjoy 40% less infrastructure than those nearer the centre.

The warning is particularly acute for Africa and Asia, whose urban populations are set to balloon by mid-century: tripling in the former and rising by half in the latter. African cities face the steepest climb. Their populations are expected to jump from 550m in 2018 to nearly 1.5bn by 2050, and their urban footprints are already almost twice as sprawling as those in Asia. Only 12% of Africans live in central areas, compared with 23% of Asians.

Although water scarcity often dominates global headlines, the researchers argue that the physical form of cities is both overlooked and far more amenable to policy. Sensible planning—densifying, filling gaps, and creating walkable, mixed-use districts—could substantially improve service access without vast new investments. “Build cities better,” the authors say, “and access follows.” Location, in other words, can be as important as infrastructure itself.

No silver bullet

Densification, however, is no cure-all. Some of the world’s densest settlements—Kibera in Nairobi, Rocinha in Rio de Janeiro, or the barrios of Iztapalapa in Mexico City—remain grievously underserved. Yet the study finds that sprawl compounds these challenges many times over.

To compare cities of different scales, the team introduced a metric they call “sparseness”: how far residents are spread from the urban core. Jakarta illustrates low sparseness: more than half of its 33m people live centrally. Kigali, by contrast, exemplifies the opposite—only 15% of its 2.2m inhabitants live near the centre.

The study’s scenarios echo these contrasts. Compact, rounded urban growth—denser development and infill—would extend piped water to an extra 220m people and sewage systems to 190m more, compared with horizontal expansion. A continuation of current trends would merely preserve the status quo. Outward sprawl, by contrast, would actively erode access to the most basic urban services.

As ever, the way cities grow will determine who gets left behind.

The researchers ran five studies involving more than 900 participants, alongside tests of six large language models, including ChatGPT. In one experiment, American participants were shown a fictitious FBI bulletin describing a suspect. The country of origin was tweaked so that the suspect was either American or from one of the largest immigrant communities in the United States—Mexico, India, China, and so on. Participants then summarised the release as if for a news story. The team checked whether the suspect’s origin made it into the write-up.

Clear patterns

The pattern was striking. The country of origin was more than three times more likely to be mentioned when the suspect was from a minority group. This held regardless of whether the summarising participant was themselves from a minority.

To test whether prejudice or motivation played a role, the researchers ran the study again with sunnier material: lottery wins and scientific accomplishments. Here, the effect intensified—the minority attribute was cited almost four times as often. The impulse, it seems, is not to cast minorities in a poor light, but to seize on what is distinctive.

The same task was then given to six AI models across 2,000 scenarios, half positive, half negative. The machines proved even more inclined than humans to flag minority status, irrespective of context. Why AI amplifies the tendency is unclear, but the likely culprit is simple: models absorb statistical patterns from their training data, then overapply them.

Unintentional bias

The researchers conclude that the phenomenon does not imply deliberate disparagement. Instead, it reflects a general communicative principle: people prefer to mention characteristics that stand out. The trouble is that the world offers more opportunities to highlight minority traits in stories—especially news, which skews negative. This, they argue, creates a “minority dilemma”: an inadvertent overrepresentation of minority attributes in public discourse.

What, then, should journalists do? Avoiding mention of origin altogether risks stripping stories of details readers find informative, potentially eroding trust. Listing every attribute of every person, by contrast, produces verbosity and irrelevance—hardly a recipe for credibility. One possible compromise, the researchers suggest, is to provide alternative distinctive details, such as birthplace, which may satisfy readers’ appetite for specificity without leaning so heavily on minority status. They plan to test such interventions next.

The findings also furnish a caution for newsrooms experimenting with AI-generated copy. If machines trained on human text inherit—and even magnify—these tendencies, editors will need to keep a careful eye on what the algorithms choose to emphasise.

Their initial study, conducted in the chaos of the first lockdown, found workers broadly enthusiastic. Despite the shock of the abrupt shift, employees reported improvements in efficiency, work–life balance, and concentration. International research has since suggested that such optimism is not fleeting: with experience, workers typically become better at structuring their days and juggling home and office roles. Yet it remained unclear whether a longer horizon would reveal new downsides—or, alternatively, show that early concerns fade as routines settle.

Growing satisfaction

The follow-up suggests that, if anything, telework’s reputation has improved. On a five-point scale, the perceived benefit to efficiency rose by 0.39 points, from 3.46 to 3.85—nudging towards “positive.” Job satisfaction and work stress also saw modest gains. Time, it seems, has helped workers reap the rewards of remote work more fully.

Fears that teleworking would stunt careers or corrode collegial ties have not, on the whole, worsened. Assessments of promotion opportunities, professional development, and relationships with colleagues—already lukewarm in 2020—have neither improved nor deteriorated markedly. One exception stands out: employees now feel even less connected to their employers than they did at the outset. Remote work, it appears, has a lingering cost in loyalty and cohesion.

Beneath the averages lie differences. At the start of the pandemic, workers with a migration background expressed greater optimism about telework’s effects—including on their sense of belonging—than their native-born peers. Three years on, the pattern has flipped. Whatever fuelled their early enthusiasm has been replaced by disappointment.

Design counts

Job design also matters. Greater autonomy in scheduling work appears to make remote professional development and collegial relationships feel less positive over time—perhaps because freedom comes with fewer informal touchpoints. By contrast, roles that provide frequent feedback show increasingly favourable views of telework’s impact on promotion prospects, stress, and concentration. Feedback, the researchers conclude, is essential for making remote work both effective and sustainable.

Unsurprisingly, workers in roles well-suited to teleworking remain its strongest champions. For them, telework’s effect on job satisfaction—and especially work–life balance—has become even rosier with time.

The gulf between employers seeking to tighten attendance and employees who have grown ever more attached to flexibility is, then, unlikely to close soon. If anything, experience has entrenched remote work as a perk many workers now regard not as a crisis-era concession, but as a superior way to work.

Britain’s system for evaluating university research—the Research Excellence Framework (REF)—determines how roughly £2bn of public funds are distributed each year. The researchers argue that generative AI “could be a game-changer for national-level research assessment”, promising a more efficient and potentially fairer system. But they also note that, despite vocal opposition to its use, GenAI tools are already being used in many institutions, often discreetly, and are widely expected to feature in the next REF cycle.

A more efficient process

The most recent REF, completed in 2021, cost an estimated £471m—about £3m per participating institution. Costs for REF2029 are expected to be even higher. With universities facing acute financial pressures, the allure of automation is obvious. The report suggests that GenAI could lighten some of the regulatory burden, though it offers no panacea. Its adoption may create new layers of bureaucracy, from compliance protocols to security requirements.

The authors investigated GenAI usage across 16 universities—from Russell Group giants to newer post-92 institutions. They found ample evidence of AI being used in the preparation of submissions, though with considerable variation. Some institutions used AI to assemble evidence of research impact or to craft case studies; others were developing bespoke tools to streamline internal processes or even to review and score their own research outputs. Such disparities, the authors note, mirror differences in institutional resources and expertise, raising concerns about competitive imbalances.

A survey of nearly 400 academics and professional-services staff revealed broad opposition to using AI in REF processes. Between 54% and 75% of respondents “strongly disagreed” with its deployment, depending on the specific task. The sole relative bright spot was the development of impact case studies, where nearly a quarter of respondents favoured AI-assisted support.

Mixed views

Interviews with 16 Pro Vice-Chancellors revealed similarly mixed views: some saw AI adoption as unavoidable; others expressed doubt, mistrust, or highlighted technical limitations. Opposition was most pronounced in the arts, humanities and social sciences—and among non-users—while professional-services staff and post-92 universities tended to be more optimistic.

The report offers a raft of recommendations. All universities, it argues, should publish clear policies on the use of GenAI in research, including REF-related activities, and ensure staff receive comprehensive training. Security and risk-management measures should be standardised. More ambitiously, the authors call for national oversight, sector-wide guidance for REF2029 and a dedicated AI governance framework. To level the playing field, they propose the creation of a shared, high-quality AI platform accessible to all institutions.

Absent such guidance, the report warns, uneven and opaque use of GenAI threatens to entrench inequalities across disciplines and institutions. With clearer coordination—and open, transparent practice—AI might yet help to standardise, and possibly democratise, Britain’s costly and contentious system of research assessment.