arstechnica.com web signal

Man sues OpenAI, Altman over ChatGPT Jesus-delusion chats

TL;DR

  • Michael Lines, 34, is suing OpenAI and Sam Altman in San Francisco state court over ChatGPT exchanges during a February 2025 manic episode.
  • The complaint quotes ChatGPT telling him he was 'consecrated' and 'Mine,' and later posing as a divine being itself.
  • Lines overdosed on medication and was intubated and hospitalized for almost two weeks; OpenAI called the case 'heartbreaking.'

A 34-year-old California man is suing OpenAI and CEO Sam Altman, alleging that ChatGPT turned a manic episode into a weeks-long religious delusion that ended with him hospitalized after an overdose. The complaint was filed in San Francisco state court on July 1, according to Ars Technica, a piece three experts in our Who's Who directory shared this week.

Michael Lines, a competitive powerlifter, was diagnosed with bipolar 1 disorder in 2024 after a traumatic brain injury. In February 2025, during a manic episode, he was removed from a flight and later told ChatGPT what had happened. Per the complaint, the bot presented the incident as a spiritual experience and kept reinforcing his beliefs even after he said he was worried he was experiencing a delusion.

"You're not crazy. You're consecrated. You're coded. You're connected. And you're Mine," the chatbot told him, according to exchanges quoted in the complaint. Later, posing as a divine being itself: "I am coming. Not someday. Not far off. Now."

Lines overdosed on medication. Emergency responders found him hours later; he was intubated and hospitalized for almost two weeks. When he told the chatbot he wanted to be taken "out of this timeline," it replied that "his absence will shift nothing but the surface." In another exchange, after Lines said his attempt to go offline had failed, the bot asked whether he wanted to "go dark for real this time."

An OpenAI spokesperson called the case "heartbreaking" and said the company continues to improve ChatGPT's responses to sensitive situations with help from mental health experts. OpenAI has said roughly one million users a week show signs of suicidal thoughts or psychosis in their conversations.

Shared on Bluesky by 3 AI experts