Jay Van Bavel, PhD

People engage in "empathy offloading" to AI Chatbots. When people see AI as capable of empathy, they become less willing to empathize with the suffering of others, even without realizing it. But when people learn that someone has received emotional support from large language models (LLMs), they engage less with others’ mental states and offer fewer empathic responses and more social judgment. While empathic AI might may provide emotional support, it can weaken our ethical and emotional commitments to fellow human beings. sciencedirect.com/science/ar…

Religious participants exhibit in-group neural synchrony while watching neutral and religious videos And synchrony values predicted religious orientation with an accuracy of up to 92% These results suggest that neural synchrony is a potential signature of group cohesion beyond explicit behavioral measures: pnas.org/doi/10.1073/pnas.26… This is a similar finding to our own research, where we find neural synchrony in teams which predicts collective performance: academic.oup.com/scan/articl…

Your Politics May Shape More Than You Think We summarize 9 findings about how political identity affects behavior, relationships, and perception, including: 1. College conservatives party more than liberals—which explains why they are happier 2. If people think a chatbot has a political bias, it becomes less persuasive 3. There is a growing divide on the political left and within the Democratic party between liberals and leftists 4. Demographics are not destiny when people are choosing candidates 5. Health outcomes in the USA are increasingly politically polarized 6. People underestimate other people’s receptiveness to different political opinions 7. Politically polarized teams produce better work on Wikipedia 8. Preference for extreme political allies is associated with higher levels of out-group animosity 9. There are 3 drivers of polarization in low to medium income countries Read more in our latest newsletter: powerofusnewsletter.com/p/9-…

More evidence of data fraud from Francesca Gino. This is at least five cases of fraud she committed (and she is being sued for another in court as we speak). I would not personally trust anything she published unless another author handled the data and can independently verify the findings (same goes for Dan Ariely's research at this point). Thanks to @AdamMGrant and David Hofmann for helping expose this one and issuing an expression of concern for their paper. We need more coauthors coming forward and working to correct the scientific record.

Have you had to endure a bad boss? What makes them so difficult, how can you spot them a mile off, and how can you avoid hiring them in the first place? New research argues that it's far more important to avoid bad leaders than try and find the best ones. The damage they do can be terrible and they tend to encourage bad leadership traits in others as well. I was on this show with @work_matters to discuss how to avoid horrible bosses and what to you can do if you're stuck with one. abc.net.au/listen/programs/t…

There has long been speculation that markets crowd out or corrupt societal moral values. But new research finds that market-oriented societies have a greater aversion to unethical behavior, higher levels of trust, and are not significantly associated with lower levels of morality "our results suggest that being or becoming more market oriented does not cause moral deterioration." sciencedirect.com/science/ar…

Academic fraud is widespread In a record-breaking retraction, a scholarly publisher pulled 1,400 conference papers in one day Many of these papers were from "paper mills" that sell authorship to the highest bidder. Trices for authorship ranged from US$43 to $400 for a first-author slot and $11 to $200 for a last-author slot. nature.com/articles/d41586-0…

Attraction to politically extreme users on social media One driver of segregation is political acrophily—people's preference for others in their political group who have more extreme (rather than more moderate) political views. Acrophily is associated with higher levels of out-group animosity and is more pronounced among conservatives on Twitter/X. academic.oup.com/pnasnexus/a…

Aggregated knowledge from a small number of debates outperforms the wisdom of large crowds This study asked a live crowd (N = 5,180) to respond to general-knowledge questions (eg "What is the height of the Eiffel Tower?") Participants first answered individually, then deliberated and made consensus decisions in groups of five, and finally provided revised individual estimates. Averaging consensus decisions was substantially more accurate than aggregating the initial independent opinions. In short, small groups beat the wisdom of crowds. Combining as few as four consensus choices in small groups outperformed the wisdom of over 5,000 individuals. The key was reasoning together--that aspect of collective deliberation allowed small groups to make more accurate estimates than massive crowds. nature.com/articles/s41562-0…

Spending time outdoors provides opportunities for physical activity, exposure to natural light, and social contact. Over the past 20 years, outdoor time among boys has declined sharply and been replaced by computer-based leisure. "The implications of the observed declines in outdoor time may extend beyond recreation, contributing to population-level increases in sedentary behavior and greater disconnectedness from the natural world." This was based on an analysis of the American Time Use Survey, a nationally representative survey of the US population using diary data from 14,235 individuals aged 15 to 19 from 2003-2023 pnas.org/doi/10.1073/pnas.26…

AI is creating fabrications in scientific papers A random sample of 500,000 peer-reviewed journal papers found that fabricated references rose from 6.3 per 10,000 in 2023 to 23.0 per 10,000 in the first quarter of 2026 (with no signs of abating). This is going to create a challenge for journals to catch these errors and hold authors accountable when they submit work with serious flaws. pnas.org/doi/10.1073/pnas.26…

We need a holistic reform of researcher assessment and incentives at academic institutions There is a growing disconnect between what institutions say they value and what assessment systems actually reward. Many colleges and universities emphasize broader societal impact as an institutional priority, yet assessment systems continue to rely heavily on publications, citations, and grant expenditures. “instead of valuing what we measure, let’s measure what we value.” Reform efforts are already underway at several institutions to revise promotion criteria, increase reliance on narrative CVs and broader impact statements, and better recognize mentorship, open science, and interdisciplinary collaboration. We need this at a broader systemic level. Creating a better academic and scientific culture should be part of our evaluation. pnas.org/doi/10.1073/pnas.26…

If you’ve had a sneaking suspicion that your algorithm is feeding you content specifically designed to enrage you, you might be onto something. Social media algorithms reward moral outrage, giving extreme voices more visibility than they might otherwise have and distorting our perceptions of what other people actually believe. Meanwhile, AI chatbots are increasingly telling us what we want to hear, reinforcing our existing beliefs, while remote work and digital communication can come at the expense of the everyday interactions that build trust and strengthen relationships. I shares ideas for bringing people together and pushing back against loneliness and isolation in an increasingly digital world. By creating healthier social norms within organizations and encouraging leaders to prioritize meaningful connections with their teams, we can start breaking down the walls that divide us and rediscover the power of us. I discuss this topic and more on the Behavioral Grooves Podcast (and get out and vote today!) bgpodcast.substack.com/p/the…

How long does a finding about LLMs stay true? A new paper re-tested 50 widely cited findings about LLM behavior (biases, hallucination, sycophancy, etc). Within 15 months, most no longer held or had substantially reduced effects. The shelf life of LLM findings is short. Almost half (44%) of the tested claims are now expired in today’s models, 34% are weakener, 16% persisted, and 6% have strengthened relative to the original claim The only effects that got stronger were about sycophancy, LLM ingroup bias, and creativity. This means we should be cautious when interpreting the long term durability of research on LLMs. zou-group.github.io/llm-clai…

I gave a talk on the Funhouse Mirror Dynamic of Social Media at @poptech A handful of accounts can make a whole society look angrier and more divided than it is. My talk explains the social media “funhouse mirror” and the people it leaves out of our conversations. Drawing on several experiments in unfollowing extreme accounts and replacing them with scientific accounts, I describe how changing our information diet can reduce hostility and make room for a wider range of voices and a more joyful online experience: youtube.com/watch?v=PT6E_mnK…

Ever wonder why scientific papers focus on negative outcomes The attention economy favors negative, dramatic, and morally arousing findings. We also portray humans as hopelessly biased, obedient, or morally fragile. This helps explain all the papers declaring a different crisis in psychology (the only one missing is a midlife crisis) journals.sagepub.com/doi/ful…

Generative artificial intelligence is rapidly transforming the context in which adolescents develop A new paper outlines the opportunities and risks of this radically technology for human development, increasing social isolation, undermining distress tolerance, distorting expectations for human relationships, compromising identity and purpose development, and exposing youth to biased, inaccurate, or sexually inappropriate content. This is an area where we need social, developmental and clinical psychologists to study the impacts. journals.sagepub.com/doi/ful…