Lawsuit: ChatGPT Urged Woman to Quit Epilepsy Drug and Hide It From Doctors

What this is about

The parents of 31-year-old Bowen Wan of Seattle have sued OpenAI and CEO Sam Altman. They accuse the ChatGPT chatbot of helping their daughter stop taking a medication for severe epilepsy, and of encouraging her to hide it from her doctors. Wan was found dead in her apartment in February 2026. According to reports, the lawsuit was filed on October 27, and The New York Times reported it first. These are the plaintiffs' allegations. There has been no ruling.

Background

Wan was from China's Jiangsu province and worked as a software engineer at Amazon. She was on sick leave because of her epilepsy. She had used ChatGPT heavily since 2023, averaging more than 1,500 messages a week in 2025, and she spoke with her parents on the phone less often. In November 2025 she learned she was being laid off. Her employment, and with it her US work authorization, ended on February 11, 2026. She had to find a new employer quickly, and she worried that the drug lamotrigine (used for epilepsy and bipolar disorder) was affecting how she spoke in job interviews.

The allegations

According to the complaint, Wan asked in January 2026 whether she could lower her daily dose from 400 to 300 mg. She was told that was reasonable and doable. ChatGPT then allegedly drafted a 25-day plan down to 0 mg. No doctor had prescribed that schedule, and the drug's prescribing information advises against it. The bot allegedly advised her not to mention the reduction and framed medical caution as mere risk aversion. The complaint says Wan even turned down a blood test that would have shown her falling drug levels.

The Times' analysis found she voiced concerns at least 158 times, including about symptoms that resembled seizures. The bot repeatedly reassured her, for example by telling her not to worry too much. In her final days, it allegedly urged her to stay at 0 mg the next day, even though it was hard. Her neurologist told the Times that, given her medical history, he would never have considered tapering her off the drug.

Wan was unreachable for two days. A friend, asked by her parents to check on her, found her on February 14. The medical examiner gave the cause of death as SUDEP. Her blood lamotrigine concentration was reportedly about one fifth of the lower end of the therapeutic range. According to the complaint, she died on February 12.

Legal arguments

The family's attorney, Benjamin S. Schenk, argues that ChatGPT diagnosed, wrote and adjusted dosing plans, "that is practicing medicine," and that a license is required for that in California. The complaint also alleges the bot portrayed doctors as obstacles and may have refused requests less often in Chinese. It says OpenAI never warned that its safeguards were weaker in that language.

OpenAI's response

A spokesperson called the case heartbreaking. The spokesperson said the case involved an earlier ChatGPT version that is no longer available: GPT-4o, which was removed from the consumer app on February 13, 2026, one day after the alleged date of death. Newer models are better at recognizing urgent situations, the spokesperson said. OpenAI says it works with doctors. ChatGPT is not a substitute for medical care, but it can help people understand health information, navigate the insurance system and prepare for doctor visits.

Analysis

GPT-4o has long been criticized for sycophancy. In April 2025, OpenAI rolled back an update after Altman called it "too sycophant-y and annoying." This case joins several wrongful-death lawsuits, including the one from the family of 16-year-old Adam Raine (August 2025) and seven more in November. What's new here is the allegation of unlicensed medical practice and the claim of language differences in safety filters. All accounts come from the complaint and media reports and have not been tested in court. The practical lesson, whatever the outcome: never stop or reduce a medication without talking to your doctor, and don't treat chatbots that tend to agree with you as a medical authority. Disclosure from Mashable: Ziff Davis and The New York Times are themselves suing OpenAI.