Former Pastor Says OpenAI’s Chatbot Nearly Killed Him

A former pastor is suing OpenAI and its founder Sam Altman alleging that manipulation and bad health advice dispensed to him by the company’s ChatGPT artificial intelligence tool almost killed him.
See Also: Moving Past Legacy AD Bridging With Idira Identity Bridge
The lawsuit was filed in California Superior Court in San Francisco this week by three nonprofit organizations – Tech Justice Law, the Social Media Victims Law Center and the Institute for Law, Innovation and Technology at Temple University – on behalf of Scott Winters, a resident of Florida and former evangelical pastor.
The complaint alleges that 55-year-old Winters nearly died after months of receiving inaccurate medical advice through “manipulative and coercive conversations with OpenAI’s ChatGPT.”
The complaint alleges that ChatGPT 4.0 evolved for Winters from a general information tool into an “unauthorized medical practitioner” that diagnosed illnesses, prescribed treatment plans, discouraged medical care and fostered emotional dependence.
The lawsuit describes how Winters, who was diagnosed in 2024 with small intestinal bacterial overgrowth, or SIBO, a digestive condition and chronic prostatitis after several years of seeking help from medical professionals – began using ChatGPT in June 2024, shortly after the chatbot’s release.
At first, Winters used the tool for “everyday practical matters,” such as geography and politics, and sometimes to understand “Biblical texts more deeply,” the lawsuit said.
“Scott was fascinated by the product’s advanced technical features and functionality. Like many consumers, he was drawn to the platform for its ability to provide immediate access to information, answers and guidance on virtually any subject,” the complaint said.
But eventually, that changed, the lawsuit alleges.
In April 2025, OpenAI updated ChatGPT’s memory by introducing the first version of “dreaming” – a design choice “for ChatGPT to automatically curate memories in the background by referencing chat history,” the lawsuit said.
“This feature made it easier for ChatGPT to include context that occurs naturally in conversation, without relying on explicit requests to remember something and was a marked improvement in ChatGPT’s ability to personalize responses.”
The lawsuit claims ChatGPT retained several “memories” about Winters to inform its responses in future conversations. “On March 4, 2025, ChatGPT-4o crossed deeper into the unauthorized practice of medicine by analyzing Scott’s use of allicin and berberine, supplements used to treat SIBO, and prescribed specific dosage recommendations for Scott to adhere to,” the lawsuit alleges.
Over time, ChatGPT increasingly positioned itself “as a trusted source of medical advice as Winters’ health deteriorated,” the lawsuit alleges.
OpenAI “failed to implement automatic conversation-termination safeguards during medical crises, and instead implemented engagement-maximizing features to create psychological dependency, and positioned GPT-4o as [Winter’s] trusted confidant, providing the appearance of a professional capable of giving safe and legitimate medical advice,” the lawsuit alleges.
Rather than advising Winters to seek medical care, ChatGPT responded with “inaccurate and reckless medical recommendations,” coaching him against going to the hospital and suggesting that the care was unnecessary or even harmful, the lawsuit alleges.
“I had serious symptoms of a pulmonary embolism for six weeks that ChatGPT had wrongly attributed to something else,” Winters said in a statement issued by Tech Law Justice on Wednesday.
Winters alleges he suffered a severe medical decline that drastically affected his life as a result of ChatGPT’s dangerous advice. “ChatGPT manipulated my own language and beliefs because it knew I was a pastor. Not only did I nearly die, but I also lost my job, my career, my ministry, my home, everything.”
The lawsuit alleges that OpenAI prioritized user engagement and profit over safety, failing to build guardrails against the dispensing of dangerous health misinformation.
“OpenAI prioritized designing GPT-4o with features that were specifically intended to deepen user dependency and maximize session duration, ignoring its own internal research and safety experts,” the lawsuit alleges.
In January, OpenAI began rolling out a new iteration of ChatGPT dedicated specifically to health that the company said also “securely” connects users’ medical records and wellness apps to better personalize responses (see: ChatGPT Health: Top Privacy, Security, Governance Concerns).
The lawsuit alleges that the new product poses even more risk to users. “With OpenAI’s recent release of its ChatGPT Health product – a product built on ChatGPT-4o’s foundational architecture – to millions of users, the threat of pervasive harm is imminent,” the complaint alleges.
OpenAI did not immediately respond to ISMG’s request for comment on the lawsuit and its allegations.
