A Florida man sued OpenAI on Wednesday, alleging ChatGPT’s medical recommendation discouraged him from in search of therapy for signs that preceded a life-threatening pulmonary embolism, practically costing him his life.
Scott Winters, a former pastor, filed go well with towards OpenAI and CEO Sam Altman in San Francisco County Superior Court docket on Wednesday, claiming the chatbot made inaccurate suggestions that imperiled his well being.
Winters argues ChatGPT crossed the road from offering info into training medication with no license, whereas OpenAI says its chatbot will not be meant to exchange medical professionals. The lawsuit claims OpenAI prioritized consumer engagement and revenue on the expense of public security, and that it failed to supply ample guardrails to guard customers from following unhealthy well being recommendation.
“ChatGPT-4o continuously opined on and evaluated Scott’s bodily situation, conveying the authority and confidence of a medical skilled rendering a scientific evaluation and therapy advice, regardless of not having the {qualifications} to take action,” the lawsuit claims.
Winters and Tech Justice Legislation, a nonprofit aiding together with his authorized protection, additionally accuse OpenAI of exploiting Winters’ non secular beliefs.
What the lawsuit alleges
In 2025, Winters turned to OpenAI’s ChatGPT-4o mannequin to attempt to perceive what was the reason for him to commonly really feel dizzy. He additionally instructed the chatbot that he was experiencing bouts of blood stress instability.
ChatGPT allegedly dismissed each units of signs as comparatively minor and prompt he keep residence and stay “recliner-bound,” the lawsuit claims. The chatbot suggested that he’d have to expertise between eight and 10 extra episodes for his situation to be thought-about severe, recommendation that Winters adopted, the allegations state.
A few weeks later, in July 2025, he suffered what the go well with describes as a “huge pulmonary embolism” resulting from blood clots that “introduced him to the brink of loss of life.”
Considered one of his docs even blamed his immobility, really useful by ChatGPT, on bringing in regards to the embolism, in line with the lawsuit.
“Scott Winters practically died as a result of ChatGPT acted like a medical authority whereas having not one of the accountability,” Matthew P. Bergman, founding legal professional of the Social Media Victims Legislation Heart, stated in a press release. “If ChatGPT have been a doctor allotting medical recommendation, it could be responsible of medical malpractice.”
The lawsuit claims that Winters additionally consulted ChatGPT the day of his medical emergency. Hours earlier than the occasion, he requested the bot if a tenderness in his groin warranted a visit to the hospital.
The chatbot drew on Winters’ spirituality in responding, telling him that “God didn’t design your physique to endlessly fail,” the lawsuit alleges.
“It turned out that the ache Scott was feeling in his groin was the onset of the pulmonary embolism that may practically kill him,” it claims.
ChatGPT not meant for medical care
OpenAI spokesperson Drew Pusateri instructed CBS Information that ChatGPT will not be designed to face in for a healthcare supplier.
“ChatGPT will not be a health care provider and will by no means be used as an alternative choice to medical care, analysis or therapy,” he stated in a press release. He famous that it’s normal for individuals to look the web for well being info and that “AI could make that have higher by serving to them discover clearer solutions, set up their questions and put together for conversations with medical professionals.”
However he stated blaming chatbots alone for mishaps associated to 1’s personal care is misguided, “and dangers getting in the way in which of individuals accessing highly effective new instruments that may help them of their well being journey.”
ChatGPT makes clear in its phrases that the product will not be meant for medical analysis or therapy. Its phrases of service state: “You settle for and agree that any use of outputs from our service is at your sole threat and you’ll not depend on output as a sole supply of fact or factual info, or as an alternative choice to skilled recommendation.”
James Grimmelann, a professor of digital and data legislation at Cornell Legislation College, instructed CBS Information that disclaimers can shield an organization from legal responsibility — to some extent.
“Disclaimers in legislation will be efficient, however in some unspecified time in the future they offer out,” he stated. “In some unspecified time in the future, the chatbot’s personal statements and actions belie any disclaimer.”
Winters’ authorized crew is in search of monetary damages, and stronger guardrails round ChatGPT allotting medical recommendation. They’ve additionally requested that ChatGPT Well being, OpenAI’s service designed to reply well being questions, be blocked from working pending an analysis of its security.
OpenAI has confronted further lawsuits associated to customers’ well being and security.
In Might, a Texas couple whose son died of an overdose after in search of out info on medication from ChatGPT, sued OpenAI alleging he’d nonetheless be alive if the corporate had not “bypassed security guards.”
