Abstract

Generative artificial intelligence (AI) has received broad criticism for its role in spreading misinformation (1–5). In its 2024 Global Risks Report, the World Economic Forum ranked AI-amplified misinformation as one of the most severe risks that the world currently faces (6). In this context, evidence for the potential positive impacts of AI is particularly welcome. On page 1183 of this issue, Costello et al. (7) report such evidence. The authors recruited more than 2000 conspiracy believers and showed that a brief but personalized conversation with an AI-driven chatbot could durably reduce research subjects’ misinformed beliefs by 20% on average. Notably, this effect persisted for at least 2 months after the intervention and was observed across a wide range of conspiracy theories. The results challenge conventional wisdom about conspiratorial beliefs and demonstrate that it is possible to counter even deeply entrenched views with sufficiently compelling evidence.