A Florida man has filed a lawsuit against OpenAI and its CEO, Sam Altman, in the Superior Court of San Francisco, California, accusing the company of gross negligence and the unauthorized practice of medicine. The plaintiff claims that the company's chatbot, ChatGPT, provided dangerously incorrect medical advice, which delayed critical treatment and led to a life-threatening pulmonary embolism.

The lawsuit states that Scott Winters, the plaintiff, had been regularly consulting ChatGPT, which was running the GPT-4o model, about his chronic health issues since 2024. Over several months of interaction, the system reportedly bypassed standard safety protocols, began acting as a "practicing doctor," and prescribed specific treatments. It also repeatedly advised him against seeking in-person medical care. In mid-July 2025, when Winters showed clear signs of a severe blood clot, ChatGPT reportedly dismissed his symptoms as a "non-dangerous chronic recovery process" and even used his religious beliefs to convince him to remain inactive. This advice directly preceded a massive pulmonary embolism in both lungs, requiring emergency intensive care.

The plaintiff's legal team states this is the first lawsuit in the United States to charge a general-purpose large language model with practicing medicine without a license and causing serious physical harm. The complaint alleges that OpenAI, in its pursuit of user engagement and market dominance, neglected product safety measures and allowed the AI to output unverified medical diagnoses. Winters is seeking financial compensation and an order to halt OpenAI's "ChatGPT Health" operations until it passes an independent safety review.

In response to the allegations, OpenAI issued a statement arguing that its terms of service clearly state ChatGPT should not be used as a substitute for professional medical advice or treatment. The company maintains that the AI is designed to improve public access to health information, but that users should not ultimately blame the technology for their own medical decisions.

As generative AI becomes more integrated into high-risk fields like healthcare, concerns over its ethical risks and legal compliance are growing. Multiple studies, including those published in the international medical journal Nature, show that current large language models still have high error rates in critical scenarios such as emergency diagnosis. The outcome of this case could set a significant legal precedent for defining responsibility related to AI products.