.png)
POST-DOC
PHILIP PARNAMETS - NYU Social Psychology Postdoctoral Fellow
Philip completed his PhD in cognitive science at Lund University and has spent two years as a postdoc with the Emotion Lab at Karolinska Institutet. His research is grounded in a broad interest in the cognitive and computational mechanisms underlying preference change, decision making and learning, especially in the moral domain. In the Social Identity and Morality lab his work focuses on dynamic models of social learning about moral agents and of moral choices generally. Philip spends his spare time creating, listening or dancing to electronic music.
Healthy Democracies
Political polarization and the spread of misinformation are widely recognized as major threats to a thriving, democratic, and pluralistic society. At the same time, the U.S. and many other countries are more polarized than ever. Drawing from fundamental research, we develop and test solutions to address these pressing challenges. For instance, we developed one of the most effective, empirically tested strategies to date for reducing anti-democratic attitudes, support for partisan violence, and partisan animosity in the U.S. This strategy involves reminding people of their common American identity while highlighting that most partisans have more in common than they realize (find a demo here).
With over 5 billion people using social media and nearly 1 billion people using AI chatbots, it is critical to understand how polarization intersects with political processes in the digital world. We study whether and how social media may contribute to democratic backsliding, for instance, by amplifying hate speech, spreading misinformation, and facilitating the coordination of anti-democratic collective action across countries. In a set of digital field experiments, we found causal evidence that social media networks can heighten animosity towards outgroups as well as exposure to misinformation. Incentivizing participants to unfollow hyper-partisan social media influencers led to reduced hostility toward their opposing political party and increased engagement with more accurate news for several months (Rathje et al., forthcoming). AI chatbots are 50% more sycophantic—that is, more likely to be overly validating and flattering—than humans. We recently ran a set of studies to understand how interacting with these chatbots affects people. We found that people who engaged with sycophantic AI chatbots, rather than more balanced ones, became more entrenched in their existing beliefs. This foundational finding raises important follow-up questions about how sycophantic AI could influence society, for instance by reinforcing echo chambers.
Beyond digital interventions, we explore whether interactive learning tools—such as role-play-based conflict resolution programs in schools and generative AI chatbots—can foster civil discourse and encourage participation in democratic processes. Additionally, Dr. Jay Van Bavel served as an expert contributor to the European Commission’s report, Values & Identities—A Policymaker’s Guide, which translates research on values and social identities into a practical toolbox for designing, communicating, and co-creating policy with citizens.