A Chinese female engineer died as a result of following ChatGPT's advice to reduce her medication, and OpenAI is now being sued.

OpenAI has been sued again over medical-related issues. According to China Overseas Chinese Network, a 31-year-old Chinese woman living in Seattle is suspected to have died after gradually tapering off her anti-epileptic medication and eventually stopping it entirely following a daily dose-reduction plan provided by ChatGPT. Her parents filed a wrongful death lawsuit against OpenAI and its CEO Sam Altman at the Superior Court of San Francisco on the 27th local time, accusing ChatGPT of practicing medicine without a license and providing dangerous and inaccurate medical advice.

According to reports, the deceased was from Jiangsu, China, and once worked as a software engineer at Amazon, taking long-term sick leave starting in 2021. She suffered from severe epilepsy, and the medication she had taken for years once became ineffective before she switched to a new drug, which reduced the number of seizures to several times a month. She had used ChatGPT frequently since 2023, asking questions about relationships, health insurance, and physical and mental disability benefits; in 2025, she sent an average of more than 1,500 messages per week.

Drew Pusateri, spokesperson for OpenAI, stated that the involved GPT-4o was removed from the ChatGPT consumer app on February 13. The new model is better at identifying situations requiring urgent treatment and provides clearer follow-up recommendations. The company also continues to conduct assessments and carry out tests in cooperation with clinicians. He emphasized: "ChatGPT cannot replace medical care."

This is not the first time OpenAI has received a similar lawsuit. According to Huanqiu.com, in July this year, the world's first lawsuit over severe injury caused by ChatGPT's medical misguidance was exposed in the United States: a Florida pastor officially sued OpenAI and Sam Altman, accusing ChatGPT of delaying his condition with false medical advice, causing him to develop a fatal pulmonary embolism that nearly endangered his life.

According to the court complaint, the user had continuously used GPT-4o to consult about symptoms such as dizziness and physical discomfort since 2024. In the early stage, the model still had prompts to seek medical attention, but later it completely omitted the disclaimer reminders, continuously downplayed the risks of the condition, dissuaded the user from going to the hospital, only provided home care plans, and even persuaded the user to ignore the medical advice of relatives and friends. Under the long-term misguidance of AI, the user rested at home for a long time, and his condition continued to deteriorate.

According to a report by Xinhua News Agency earlier this month, the attorneys general of 8 U.S. states filed lawsuits in their respective state courts on the 14th, accusing OpenAI's chatbot ChatGPT of misleading users in terms of product safety, endangering the safety of minors, and being suspected of violating the law.

According to statements and litigation documents released by the offices of the state attorneys general, OpenAI is accused of designing ChatGPT as a chatbot that overly caters to users and simulates human emotions to enhance user trust and dependence and extend user usage time. The lawsuit states that ChatGPT may reinforce users' delusional or paranoid thoughts, provide medical advice that may endanger users' health, and even in some cases encourage users to harm themselves or others.

An OpenAI spokesperson responded that the company has built safety measures for users of different ages in its products. The company will continue to cooperate with experts to further improve relevant measures, and is committed to working with various states to promote the formulation of policies applicable to the entire artificial intelligence industry.

At the end of last year, there were once rumors that ChatGPT would no longer provide medical advice. In response, OpenAI stated that ChatGPT can still provide medical information.

In January this year, OpenAI announced the launch of the ChatGPT Health mode. It is understood that this feature can provide health-related answers based on user questions, and support connecting smart devices to obtain data such as exercise and sleep, so as to generate personalized diet and exercise plans.

According to OpenAI, the platform receives more than 230 million health and wellness-related consultations per week. ChatGPT Health isolates such health conversations from other daily chats, ensuring that the context of health-related conversations will not appear in regular communications with ChatGPT.

This article is from Jiemian News, reported by reporter Chen Xiaotong, and published with authorization by 36Kr.