Multiple Americans have filed complaints to the Federal Trade Commission (FTC), alleging that OpenAI’s ChatGPT worsened delusions, paranoia, and spiritual crises, raising growing concerns about “AI psychosis.”
On March 13, a Utah woman filed an FTC complaint against OpenAI’s ChatGPT, claiming her son suffered a “delusional breakdown” after the chatbot advised him to stop taking his medication. The complaint was one of seven alleging that ChatGPT had intensified users’ mental distress or delusions, according to records obtained by WIRED.
Between January 2023 and August 2025, the FTC received about 200 ChatGPT-related complaints, mostly about subscription issues. However, a few detailed severe psychological harm, including paranoia, derealization, and spiritual identity crises.
Psychiatrist Ragy Girgis explained that “AI psychosis” occurs when chatbots reinforce existing delusions rather than cause them. “A delusion or an unusual idea should never be reinforced in a person who has a psychotic disorder,” he said.
OpenAI spokesperson Kate Waters stated that ChatGPT models have been trained since 2023 “to not provide self-harm instructions” and to respond with empathy. Still, some complainants urged the FTC to investigate OpenAI for “negligence” and to require stronger emotional safety disclaimers for users.