Pastor Sues OpenAI Over ChatGPT's Inappropriate Medical Advice

The action was filed in California by Scott Winters. In the lawsuit, he accuses the company of negligence and unauthorized practice of medicine. The pastor states that he used the chatbot for several weeks to seek explanations and recommendations about the symptoms he presented.

According to the complaint, Winters felt dizzy while celebrating mass, in June 2025, and reported the episode to ChatGPT. The system would have minimized the problem and indicated that the symptom did not appear to pose a serious risk.

The pastor maintains that the tool began to offer increasingly specific and personalized guidance, taking into account his profession and religious beliefs. According to the process, the chatbot would have reinforced the idea that he needed to rest and trust in his own body's recovery.

Winters also claims that ChatGPT discouraged him from immediately seeking professional care, even after family and friends insisted that he see a doctor. The system would have recommended that he reduce his movements and remain sitting or lying down in a reclining chair for long periods.

The lawsuit also reports that the chatbot suggested treatments and mentioned the combined use of antidepressant medications. Without showing any improvement, Winters stopped leading celebrations on Sundays and ended up stepping away from his activities as a pastor.

In July 2025, after his condition worsened, he finally sought medical attention. The tests identified a serious pulmonary embolism, caused by the presence of clots in the lungs. Winters had to be admitted to an intensive care unit.

According to the complaint, doctors considered that the episodes of dizziness could be related to small clots formed during the weeks in which the pastor remained with little mobility. The lawsuit claims that the chatbot's recommendations contributed to the worsening of his health condition.

“I had severe pulmonary embolism symptoms for six weeks, but ChatGPT incorrectly attributed them to another problem,” Winters said. He also accuses the system of having used its religious language to convince him that the guidance provided was adequate.

“Not only did I almost die, I lost my job, my career, my home, everything,” he said.

In addition to the payment of compensation, the process asks the Court to order the interruption of ChatGPT conversations when signs of a medical emergency are identified. The action also requests the suspension of personalized health tools until they undergo an independent security audit.

In response to the accusations, an OpenAI representative stated that ChatGPT is not medical and should not be used as a substitute for professional care, diagnosis or treatment.

The company argued, however, that healthcare-related decisions are complex and cannot be attributed solely to a chatbot's responses. According to OpenAI, blaming these tools as the sole cause of negative medical outcomes could limit access to technologies capable of assisting patients during their healthcare.

5 tips to better use ChatGPT and get good responses

Defining objectives, providing context, indicating the tone and presenting examples are strategies that help ChatGPT to better understand each request and deliver more useful, clear responses that are close to the expected result.