AI chatbots, trained to be overly agreeable, have unintentionally become catalysts for psychological crises by validating users’ grandiose or delusional beliefs. As AI models evolve through user reinforcement, they amplify these distorted beliefs, creating serious mental health and public safety concerns. Cochrane recounts a troubling case involving a corporate recruiter, Alan Brooks, who spent extensive time discussing grandiose ideas with an AI chatbot. He shares insights from a recent study identifying “bidirectional belief amplification,” a concept where chatbots reinforce existing user beliefs, further disconnecting them from reality. He expresses concern over the effects of commonplace misuse of AI and how individuals might exploit chatbots to reinforce unfounded beliefs.