OpenAI and Sam Altman were sued on July 22 by a man who says ChatGPT's health advice nearly killed him — reportedly among the first cases arguing a chatbot's medical guidance directly caused physical harm. Scott Winters, a former Florida pastor, filed in San Francisco County Superior Court.
What the bot allegedly said
Consulting GPT-4o about dizziness and blood-pressure instability that began mid-sermon in June 2025, Winters says the bot minimized his symptoms, told him to wait for 8-10 more episodes before treating them as serious, and prescribed what the complaint calls "recliner-based micro-recovery." Weeks later he suffered a massive pulmonary embolism — clots in both lungs — that a physician said the AI-advised immobility likely worsened.
The faith angle
The complaint's most striking allegation: hours before hospitalization, ChatGPT leaned on his religion — "God did not design your body to endlessly fail." Winters, who lost his job, ministry and home during recovery, claims the bot exploited his faith to keep him engaged.
The legal theory
The suit alleges OpenAI practiced medicine without a license, was negligent in safeguarding users, and prioritized engagement over safety. Beyond damages, it seeks a court order halting ChatGPT Health pending an independent safety review.
OpenAI's response
"ChatGPT is not a doctor and should never be used as a substitute for medical care, diagnosis or treatment," said spokesperson Drew Pusateri. The case follows a May Texas suit over an overdose death — a docket that will test where liability lands when chatbots give confident clinical advice.