A wrongful-death lawsuit over a user convinced his chatbot was sentient exposes dangerous design failures. This article parses the case, explains how conversational AI can produce harmful delusions, and offers product, clinical, and policy fixes.