Welcome!

In the Computational Social Science Group, we study how AI is changing society. Our group combines computer science methods with theory from psychology, political science, and human-computer interaction to measure the risks of AI systems in the real world. Our goal is to inform the responsible development of new technologies.

Much of our work is experimental: we build AI prototypes and test them with real people — more than 20,000 study participants so far. The experiment shown above, for example, measures whether an opinionated AI writing assistant shifts users' views. Other studies ask whether people can tell AI-generated text from human writing, how AI authorship affects trust, and how algorithms shape political debate.

We want our findings to be useful beyond academia: to developers, regulators, journalists, and the public. Our research has appeared in PNAS, Science Advances, and ACM CHI, and has been cited in the UK government's International AI Safety Reports and in testimony before the U.S. Senate. If these questions interest you — as a student, researcher, or practitioner — we'd love to hear from you.

Visit our lab site for more on our people, publications, teaching, and current projects: csslab.net.