Monthly Archives: May 2026

AI uptake skewed toward the affluent, study highlights growing inequality risks

The rapid spread of artificial intelligence (AI)—particularly as it becomes embedded in everyday tools and services—is giving rise to a new and distinct form of digital inequality.

This concern is highlighted by communication researcher Professor Sai Wang and colleagues at the Hong Kong Baptist University, who analysed data from more than 10,000 Americans to understand better how people engage with AI. Their findings, published in Information, Communication & Technology, point to widening gaps in awareness, familiarity and use of the technology across different social groups.

The study shows that individuals with higher levels of education or income are more likely to recognise AI in action, feel knowledgeable about it, and actively use it. The researchers define awareness as the ability to identify where AI is being used. At the same time, familiarity reflects a person’s perceived understanding of the technology, regardless of whether that knowledge is accurate. These differences mean that more advantaged groups are better positioned to benefit from AI.

Professor Wang stresses that gaps in awareness can have tangible consequences. If only certain groups understand how AI operates, they can use that knowledge to their advantage. For example, job applicants who know that employers rely on AI screening tools can tailor their applications accordingly, while those who lack this awareness may miss opportunities without realising why.

Awareness also shapes how individuals respond to the risks associated with AI. Those who are more informed may be better able to recognise misinformation, understand issues such as deepfakes, and navigate digital environments more critically. By contrast, individuals with lower awareness are more vulnerable to being misled or manipulated by AI-driven content.

Interestingly, the researchers found that perceived familiarity with AI was a stronger predictor of awareness than actual usage. This may be because many AI systems are seamlessly integrated into everyday platforms, making them less visible. For instance, recommendation algorithms on services such as Netflix and Spotify personalise content using AI, yet many users do not recognise these features as such.

This hidden presence of AI distinguishes emerging inequalities from traditional digital divides, which have focused on access, skills and outcomes. Because people often interact with AI without realising it, improving access alone is unlikely to close the gap. Instead, the researchers call for targeted efforts to improve AI literacy—through education, outreach and practical guidance—to help ensure that the benefits of these technologies are shared more equitably and do not further deepen existing social inequalities.

More information: Sai Wang et al, Socioeconomic disparities in AI awareness: examining the mediating roles of AI usage and familiarity, Information, Communication & Technology. DOI: 10.1080/1369118X.2026.2652505

Journal information: Information, Communication & Technology Provided by Taylor & Francis Group

Restrictions on Trade Secrets Could Weigh on Wages Near Retirement

Leave a Reply

Policies designed to safeguard trade secrets and stimulate innovation may carry unintended consequences for workers later in their careers. A new study published in Labour Economics, led by researchers from Penn State, the Federal Reserve Bank of Cleveland and Colorado State University, finds that such measures can suppress wage growth over time and encourage firms to rely more heavily on automation rather than human labour.

“The policy is often framed as a way to protect innovation, but the data tell a different story,” said Bharadwaj Kannan, associate clinical professor of finance at Penn State’s Smeal College of Business. According to the study, younger workers may benefit initially through higher starting salaries, but this comes with a trade-off: slower wage progression in later years. At the same time, firms appear to be quietly shifting investment away from labour and towards capital-intensive production.

To better understand these dynamics, the researchers examined how companies that depend on trade secrets respond to policies that restrict labour mobility—meaning how easily employees can move between competing firms. Their analysis focused on the adoption of the inevitable disclosure doctrine (IDD). This legal principle allows employers to prevent workers from joining competitors if there is a risk that trade secrets could be disclosed, even without formal non-compete or non-disclosure agreements. Proponents argue that such policies strengthen research and development and support long-term firm growth.

The team analysed firm-level financial and accounting data drawn from the CRSP/Compustat Merged database, alongside industry-level measures of employment and capital investment from the U.S. Bureau of Economic Analysis. Their dataset covered the period from 1977 to 2011 across all 50 U.S. states. Over that time, 21 states adopted the IDD at some point, while six later reversed course, leaving 15 states—including Pennsylvania—recognising the doctrine by the end of the study period.

The findings challenge the policy’s intended benefits. Rather than boosting research and development or overall firm expansion, the adoption of IDD was associated with a shift in compensation patterns and production strategies. Workers at the beginning of their careers tended to receive higher initial pay, but their wage growth slowed significantly as they aged, effectively redistributing earnings across the life course.

At the same time, firms increased their reliance on capital. Investment in equipment and machinery rose by 3.5%, while the capital-to-labour ratio increased by 5.5%, signalling a measurable tilt towards automation. Taken together, the results suggest that policies aimed at protecting intellectual assets may reshape labour markets in subtle but important ways, with long-term implications for earnings trajectories and the balance between human and machine work.

More information: Bharadwaj Kannan et al, Replacing labour with capital: Evidence from aggregate mobility shocks, Labour Economics. DOI: 10.1016/j.labeco.2025.102832

Journal information: Labour Economics Provided by Penn State

AI Delivers Best Results When Paired with Human Insight

Leave a Reply

A new academic study suggests that artificial intelligence is most effective when used to enhance human thinking and decision-making rather than replace it. Published in the Journal of Knowledge Management, the research explores how AI and human intelligence can work together to improve how knowledge is created, interpreted, and shared across society.

The authors, from the University of East London, explain that while AI is highly effective at processing large volumes of information, identifying patterns, and generating insights, human involvement remains essential for interpreting results and making ethical, context-sensitive decisions. Drawing on findings from 90 studies published since 2015, the paper concludes that the strongest outcomes arise when AI’s speed and scale are combined with human creativity, critical thinking, and oversight—enhancing problem-solving across fields such as healthcare, education, management, and research.

The study also cautions against over-reliance on AI, stressing the need for clear safeguards to ensure systems are transparent, fair, and subject to human review, particularly in high-stakes environments. Co-author Dr Susan Akinwalere noted that the true promise of AI lies in helping people navigate complexity more efficiently, while leaving judgment, meaning, and responsibility firmly in human hands.

Researchers from the Royal Docks School of Business and Law argue that this human-centred approach can help organisations build stronger “knowledge ecosystems”. Rather than treating AI as a substitute for expertise, they recommend designing it as a collaborative partner that supports learning, innovation, and informed decision-making, with appropriate checks in place to ensure context, fairness, and integrity.

More information: Susan Akinwalere et al, The symbiotic roles of artificial intelligence and human intelligence in advancing knowledge ecosystem, Journal of Knowledge Management. DOI: 10.1108/JKM-03-2025-0362

Journal information: Journal of Knowledge Management Provided by University of East London

A Novel Scale for Assessing Evidence Use in Evidence-Based Management is Developed and Validated

Leave a Reply

Organisations increasingly adopt evidence-based management to inform decision-making, yet empirical research examining how it is practised remains relatively limited. A new study addresses this gap by developing and validating a purpose-built measurement tool—the Evidence-Based Management Source Utilisation Scale (EBM-SUS). This scale is specifically tailored to management settings and captures the extent to which decision makers draw on four distinct sources of evidence.

The research, conducted by scholars from Carnegie Mellon University and the University of Malta, is published in the International Journal of Organisational Analysis. It contributes to a growing body of work seeking to operationalise and measure evidence-based practices beyond clinical or healthcare contexts.

“Although researchers have advanced understanding and highlighted the importance of evidence-based management, relatively few studies have examined how different sources of evidence are used in decision making,” says Denise M. Rousseau, professor of organisational behaviour and public policy at Carnegie Mellon’s Heinz College and a coauthor of the study. She notes that the research was designed to address this important gap in the literature.

Evidence-based management is widely recognised as a critical approach in organisational settings, strengthening decision-making through the integration of four foundational sources: scientific research, organisational data, professional expertise, and stakeholder perspectives. Each of these contributes unique insights, yet existing measurement tools often prioritise clinical forms of evidence and do not readily transfer to broader organisational contexts.

To address this limitation, the researchers developed the EBM-SUS. They evaluated it through two empirical studies involving senior leaders in the Maltese public service, including director generals, managers, and permanent secretaries. These roles involve responsibility for policy development, resource allocation, programme implementation, and both strategic and operational decision-making across ministries and government entities.

The first study used exploratory factor analysis to identify the scale’s underlying structure, while the second applied confirmatory factor analysis to validate the measurement model and examine its associations with traits such as risk aversion and conscientious decision making. The results indicate that the EBM-SUS demonstrates strong psychometric properties, supporting its reliability and validity as a tool for assessing how decision makers use multiple sources of evidence.

“Assessing how managers draw on each source—both individually and collectively—can enhance transparency, accountability, and effectiveness, while improving alignment between decisions and organisational goals,” explains Frank Bezzina of the University of Malta, a coauthor of the study. Vincent Cassar adds that the findings reinforce the value of evidence-based management as a framework for understanding decision-making dynamics, while also highlighting the EBM-SUS as a practical tool to support more evidence-informed organisational practices.

More information: Frank Bezzina et al, Evidence-based management in practice: measuring the use of four core sources of evidence, International Journal of Organizational Analysis. DOI: 10.1108/IJOA-11-2025-6185

Journal information: International Journal of Organizational Analysis Provided by Carnegie Mellon University

Why Cooperation Breaks Down Over Time, According to New Study

Leave a Reply

A new study from Saïd Business School, Universidad de Santiago de Chile, and William & Mary suggests that cooperation does not simply fade in a slow, predictable way. Instead, it rises and falls in repeated cycles, driven more by shifting human behaviour than by purely strategic thinking or financial pressures. The findings challenge long-standing assumptions about how and why people choose to work together over time.

“In everyday life, people often face a simple decision: contribute their share or let someone else carry the load,” explains Nicholas Sabin, the study’s first author and an Associate Professor of Behavioral Science at the Universidad de Santiago de Chile. He notes that the research offers new insight into why cooperation is so fragile, even when it benefits everyone involved. More importantly, it may help identify ways to sustain cooperative behaviour over longer periods, particularly in situations where ongoing collaboration is essential.

The research draws on a five-year field study conducted with a microfinance institution in Sierra Leone, tracking more than 47,000 repayment transactions from over 7,000 borrowers. Participants operated within a joint-liability lending system, meaning that each group had to ensure full repayment of its loan each month collectively. If even one member failed to contribute, the entire group risked losing access to future credit. This real-world setting provided a rare opportunity to observe cooperation under meaningful economic and social stakes.

According to co-author David Klinowski, an Assistant Professor of Economics at William & Mary, group lending offers an unusually realistic environment for studying cooperation. Unlike laboratory experiments, where outcomes are often abstract or low-stakes, the consequences here are tangible and immediate. Borrowers who fail to cooperate can face serious financial and social repercussions, making their decisions more reflective of real-life behaviour rather than controlled experimental responses.

The researchers observed a consistent pattern across borrower groups. Cooperation tended to start at high levels, as participants were motivated and attentive to their shared responsibilities. Over time, however, this commitment gradually weakened as individuals became fatigued, less vigilant, or more tempted to rely on others’ efforts. Each time a new loan cycle began, cooperation surged again, boosted by renewed awareness of the consequences of failing to contribute. Yet with each cycle, this rebound became shorter-lived, and cooperation declined more quickly than before. The authors describe this repeating pattern of erosion and recovery as a “punctuated decline.”

These findings run counter to traditional theories that frame declining cooperation as a result of rational learning or strategic adjustment. Instead, the study points to behavioural fatigue and shifting motivation as the primary drivers. Interviews with microfinance clients and staff reinforced this interpretation, revealing how emotions, habits, and social dynamics shape decisions to cooperate or defect. The implications extend far beyond lending groups. Whether in healthcare decisions like vaccination, civic participation such as voting, or international agreements between countries, sustained cooperation depends not only on incentives but on maintaining motivation. The researchers suggest that organisations can support long-term cooperation by introducing periodic resets, fostering intrinsic motivation, and designing systems that make cooperative behaviour easier to sustain over time.

More information: Nicholas Sabin et al, Punctuated decline of human cooperation, Nature. DOI: 10.1038/s41586-026-10380-3

Journal information: Nature Provided by University of Oxford

98% of Sustainability Messaging in Meat and Dairy Industry Flagged as Potential Greenwashing

Leave a Reply

The overwhelming majority of environmental claims made by the animal agriculture sector amount to misleading “greenwashing”, often built on vague commitments or distant projections, according to a study published on April 22, 2026, in PLOS Climate. The research was led by Maya Bach and Jennifer Jacquet from the University of Miami, alongside their collaborators.

The meat and dairy sector is responsible for a substantial share of global emissions, contributing 57% of emissions from food production and at least 16.5% of total greenhouse gas output worldwide. In this analysis, the researchers examined environmental messaging from 33 of the world’s largest meat and dairy corporations to determine whether their claims presented credible, actionable pathways for reducing environmental harm—or whether they instead fell into the category of greenwashing, defined as deceptive or intentionally misleading communication.

In total, the team reviewed 1,233 environmental claims sourced from publicly available sustainability reports and company websites, covering the period from 2021 to 2024. Of these, 841 claims (68%) were classified as climate-related, meaning they addressed greenhouse gas emissions or climate impacts either directly or indirectly. This emphasis suggests that climate change has become the dominant lens through which companies frame their sustainability narratives.

A significant portion of these claims—467 (38%)—consisted of unverifiable future-oriented statements, such as pledges to achieve carbon neutrality or restore large volumes of water by 2030. Despite the volume of claims, supporting evidence was limited: only 356 (29%) were backed by company-provided documentation, and just three claims cited peer-reviewed scientific evidence, two of which related to climate issues. Meanwhile, the number of companies committing to net-zero targets has grown, rising from four in 2020 to 17 in the current analysis. However, these commitments appear to depend largely on carbon offsetting rather than direct emissions reductions.

Using an established framework to evaluate greenwashing, the authors concluded that 98% of the claims—1,213 in total—could be classified as misleading. Examples include long-term pledges such as achieving “net climate-neutral dairy” by 2050, which often lack clear implementation strategies or measurable benchmarks.

The researchers emphasise that such patterns are not unique to the meat and dairy industry, though the sector’s environmental footprint makes the implications particularly significant. According to Bach, greenwashing is widespread across corporate sustainability reporting in this space, creating a false impression of meaningful climate progress. She warns that such messaging can mislead consumers, shape public perception, and ultimately weaken pressure on policymakers to enact substantive climate policies.

Jacquet adds that while it is unsurprising for meat and dairy companies to focus on climate issues—given the relatively high emissions associated with animal-based foods—the lack of concrete evidence and investment behind many of these claims raises concerns. When sustainability messaging is dominated by vague promises rather than measurable action, it risks functioning more as a public relations strategy than a genuine commitment to environmental stewardship.

More information: Maya Bach et al, Environmental claims, climate promises, and ‘greenwashing’ by meat and dairy companies, PLOS Climate. DOI: 10.1371/journal.pclm.0000773

Journal information: PLOS Climate Provided by PLOS

Odometer Blind Spots Can Be Expensive for Buyers

Leave a Reply

Think you are shopping wisely for a used car? New research from the McCombs School of Business at the University of Texas at Austin suggests that even careful buyers may be swayed by something as simple as the first digit on the odometer when judging a vehicle’s value. This subtle influence can shape perceptions of worth in ways that are not entirely rational, leading consumers to make decisions they might not otherwise make if they considered the full picture.

At the heart of this behaviour is left-digit bias, a well-documented psychological tendency. People often perceive £1.99 as significantly cheaper than £2.00, even though the difference is negligible, because their attention is drawn disproportionately to the leftmost digit. This cognitive shortcut simplifies decision-making but can distort judgment. When applied to car mileage, the same bias can lead buyers to overvalue vehicles that appear to fall just below a round-number threshold.

This inattention can prove costly. According to research led by marketing professor Raghunath Rao, when a car’s odometer reads just under a multiple of 10,000 miles—for instance, 49,999 instead of 50,000—buyers may be willing to pay notably more than the car is objectively worth. On average, this premium can reach as much as $170 per vehicle. Rao describes this extra amount as a kind of “mental tax” that consumers unknowingly impose on themselves by failing to assess mileage more carefully.

Although private sellers may not consistently take advantage of this tendency, dealerships are often more attuned to it. The study suggests that professional sellers are better positioned to recognise and capitalise on consumer biases, effectively capturing that “mental tax” as additional profit. In this sense, the marketplace does not simply reflect value; it also reflects how value is perceived, particularly when buyers rely on quick heuristics rather than detailed evaluation.

To explore this phenomenon, Rao and his collaborator Andreas Kraft analysed Department of Motor Vehicles data covering 4.8 million used car transactions in Texas between 2014 and 2021, representing roughly 10% of all such sales in the United States. They examined pricing and sales patterns around mileage thresholds such as 20,000, 50,000, and 100,000 miles. Their findings showed a consistent pattern: cars just below these cut-offs tended to sell at higher prices and more quickly than nearly identical cars just above them. The effect was especially pronounced at dealerships, where the bias appeared roughly twice as strong as in private sales, and was most evident around the 100,000-mile mark.

Importantly, the implications of left-digit bias extend beyond mileage alone. Similar patterns may influence how consumers evaluate other product features, such as the battery range of electric vehicles. A car advertised with a range of 310 miles may seem far more appealing than one with a range of 280 miles, even if the practical difference is relatively modest. The broader lesson is not directed at sellers but at buyers: awareness of these unconscious biases is essential. By consciously considering all available information rather than focusing on a single prominent digit, consumers can make more informed and financially sound decisions in any major purchase.

More information: Andreas Kraft et al, Market Effects of Inattention: Theory and Evidence from Left-Digit Bias, Journal of Marketing Research. DOI: 10.1177/00222437261427475

Journal information: Journal of Marketing Research Provided University of Texas at Austin