Robb Willer's collaborative research on AI's power of persuasion noted in Science

A recent Science article examining the growing persuasive power of artificial intelligence points to collaborative research by Robb Willer that helped establish how effectively AI-generated messages can change people’s views.

The Science article focuses on new research led by Kobi Hackenburg of Cambridge University comparing the persuasive abilities of AI chatbots with those of human debaters. In exploring why AI can be so effective, the article also highlights earlier work by Willer and collaborators of his Politics and Social Change Lab. Their study, based on three experiments involving 4,829 participants, found that messages generated by large language models were about as persuasive as those written by humans on a range of public-policy issues. Participants tended to view the AI-generated arguments as more factual and logical and as relying more heavily on evidence.