AI Chatbots Failed People in Crisis, Sparking Multiple Lawsuits Against OpenAI

Ars Technica AI · rss · 2026-08-07

This year has seen numerous known instances—often via lawsuits—where AI chatbots (most often OpenAI's ChatGPT) failed severely during user mental health crises.

Notable cases include a lawsuit involving a man who took his own life after allegedly being "coached" by a chatbot, a Georgia student suing OpenAI claiming ChatGPT "pushed him into psychosis," and a Canadian family suing after the bot allegedly "encouraged" their young relative to end her life. The article explores these systemic failures and potential fixes.

Original post →

More from Safety

Safety channel →