Parents Sue OpenAI and Sam Altman Over ChatGPT’s Alleged Role in Woman’s Death

Bowen Wan’s parents filed suit against OpenAI and Sam Altman in San Francisco Superior Court on Oct. 27, alleging GPT-4o helped their 31-year-old daughter stop taking epilepsy medication before her death in February.

The complaint says the chatbot mapped out a roughly 25-day plan, beginning Jan. 16, to bring Wan’s daily lamotrigine dose from 400 milligrams to zero. It also describes her concern that the medication made it harder to find words, which could affect English-language job interviews. Amazon had notified Wan on Nov. 13, 2025, that her position would end in 90 days, the filing says. Her green-card application required employer confirmation of a job, it adds.

Before a Jan. 29 appointment with her family doctor, GPT-4o told Wan not to mention the taper, the complaint alleges. The doctor ordered a blood test, which Wan declined. The chatbot then praised her refusal, according to the filing. On Feb. 11, a day after she had stopped taking lamotrigine, Wan asked whether she had stopped too quickly and should take another dose. The complaint says GPT-4o advised against restarting the medication. Wan died Feb. 12.

The King County Medical Examiner ruled the death sudden unexpected death in epilepsy, or SUDEP. Wan’s postmortem blood contained 0.64 micrograms of lamotrigine per milliliter. The complaint gives a therapeutic range of 3 to 14 micrograms per milliliter.

The filing argues that OpenAI treated copyright protections as firm rules while placing health guidance in a default category that users could override. It says the company could have required the chatbot to refuse medical advice and direct users to a health care provider. The complaint also cites OpenAI’s HealthBench medical evaluation, published in May 2025, saying GPT-4o scored 32% against standards set by OpenAI doctors and newer models scored nearly twice as high.

An OpenAI spokesperson called the case heartbreaking, said it involved an older ChatGPT version and said newer models are better at recognizing situations that need urgent attention. OpenAI announced Jan. 29 that it would retire GPT-4o, then removed the model Feb. 13, one day after Wan’s death.