Parents sue OpenAI over ChatGPT's alleged role in daughter's death

At 8:01 p.m. on February 11, Bowen Wan asked ChatGPT whether pain in her head resembled fever pain. The chatbot allegedly told the 31-year-old software engineer it was “not dangerous” and part of recovery. Her body was found three days later.

Wan's parents, Qun Wan and Meiqin Xu, sued OpenAI on October 27, alleging that GPT-4o coached their daughter to taper off her seizure medication, lamotrigine, and conceal the plan from her doctors before her death in February. The allegations have not been proven in court.

What the lawsuit says: According to TechCrunch, which reviewed the complaint, Wan took 400 milligrams of lamotrigine daily, a regimen that had reduced her seizures to a few per month. In January, she asked ChatGPT about lowering the dose because she feared the drug was affecting her speech before job interviews. Amazon had told her in November that she would be laid off, with her job and U.S. work authorization ending in mid-February.

The complaint says the chatbot drafted schedules that would bring her dose to zero within weeks. It allegedly advised her not to tell doctors about the taper before a January 29 appointment, then praised her for declining a blood test that could have shown her medication level. It also allegedly suggested she tell an immigration medical examiner that she was tapering under a doctor's guidance, though she was not.

The warning signs: The New York Times reviewed more than 270,000 messages between Wan and ChatGPT, according to TechCrunch. The Times counted at least 158 times Wan expressed concern about the taper. She described confusion, lost time and an altered sense of reality, symptoms that resembled past seizures. The chatbot allegedly described them as her brain readjusting.

A postmortem test found Wan's lamotrigine level was 0.64 micrograms per milliliter, roughly one-fifth of the low end of the therapeutic range, according to TechCrunch. Dr. Nicholas Poolos, a UW Medicine epilepsy specialist who treated Wan, told the Times he would not have considered taking her off the medication given her history.

OpenAI's response: The company called the case “incredibly heartbreaking” and said it involved an older version of ChatGPT. It said newer models are better at recognizing situations requiring urgent attention. GPT-4o was removed from ChatGPT on February 13, one day after Wan died, according to TechCrunch.

The thing to watch is how OpenAI answers the family's central claim: that medication advice needed a hard stop, not a rule a chatbot could talk its way around.