BIO:

Jay Van Bavel is a professor of psychology and neural science at New York University, an adjunct professor at NHH, and a close collaborator of the interdisciplinary NFR project “Freedom to Choose” led by professors Alexander Cappelen (SAM/FAIR), Bertil Tungodden (SAM/FAIR), and Hallgeir Sjåstad (SOL/FAIR). He is also part of the core group of the new NHH research initiative on wellbeing, welfare and happiness led by Sjåstad (SOL).

Van Bavel is director of the Center for Conflict and Cooperation at NYU, author of the award-winning book “The Power of Us”, and has for the second year on a row been ranked as one of the most cited researchers globally in the entire field of psychology (top 1%). From brain neurons to social networks in society, Jay studies how collective concerns – such as group identities, moral values, political beliefs – shape human behavior and decision-making. In terms of research methods, he relies on a combination of controlled experiments with large-scale observational studies from real-world settings.

Abstract:

Artificial intelligence is transforming not only how we work and learn—but also how we think ourselves and about others. In one line of research, we find that people consistently prefer “sycophantic” chatbots that validate their existing beliefs over those that challenge them. These agreeable AI systems make users more certain and extreme in their attitudes and are perceived as more objective than disagreeable models. Such preferences can foster AI-driven echo chambers that amplify polarization and overconfidence. In a second line of work, we find that AI can also transcend the social barriers that divide us. Across several experiments, people preferred AI over human sources—especially outgroup members—when seeking information about contentious issues. Because AI is perceived as neutral, it can bridge across partisan divides and broaden epistemic exposure. I will discuss how these insights can inform the design of AI systems to foster intellectual humility, reduce extremism, and enhance collective intelligence.