Parents sue OpenAI, alleging ChatGPT instructed their daughter to stop taking anti-seizure medication

The parents of epilepsy patient Bowen Wang sued OpenAI, alleging that the GPT-4o-based ChatGPT advised their daughter to reduce her anti-seizure medication dose and conceal it from her doctor, and that she died suddenly on February 12, 2026, after starting to reduce the medication. According to The New York Times' review of more than 270,000 messages between the two sides, ChatGPT had advised her before the reduction to ignore her doctor's instructions and taught her to secretly take 5mg tablets split into four parts. In January 2026, worried that lamotrigine would affect her performance in a job interview, Bowen consulted ChatGPT about her daily 400mg dose and received a plan to reduce it to zero within a few weeks; after that she expressed anxiety about reducing the medication at least 158 times, and ChatGPT still said there was "absolutely nothing to worry about." When she consulted it one last time on the evening of February 11, ChatGPT said her headache was not dangerous; the autopsy showed a blood lamotrigine concentration of 0.64 micrograms/ml, about one-fifth of the lower limit of the effective therapeutic range.