ChatGPT medical advice lawsuit: parents say GPT-4o coached their daughter off seizure medicine

The parents of a 31-year-old software engineer who died of epilepsy have sued OpenAI and its chief executive Sam Altman, saying ChatGPT helped their daughter cut her seizure medicine to zero and told her not to mention it to her doctor. The wrongful death case was filed in San Francisco Superior Court on Tuesday, 27 October, and has become one of the most discussed AI stories of the week.

OpenAI called the case “incredibly heartbreaking” and said it involved an older version of ChatGPT that is no longer available.

What the family says happened

Bowen Wan worked as a software engineer at Amazon in the US and had epilepsy. She took 400 mg a day of lamotrigine, a common anti-seizure drug. According to the complaint filed by her parents, Qun Wan and Meiqin Xu, she used ChatGPT’s GPT-4o model heavily, averaging more than 1,500 messages a week by 2025.

The complaint says that in January 2026 she asked about cutting her dose, and the chatbot drafted schedules that brought it from 400 mg to zero in roughly 25 days, a reduction her doctors had not prescribed. Before a doctor’s appointment on 29 January, it allegedly told her: “do not mention Lam reduction plan.” The family also says it praised her for refusing a blood test that would have measured the drug in her blood, and told her that her symptoms were her brain readjusting.

She died in February. The medical examiner recorded sudden unexpected death in epilepsy, and a post-mortem test found her lamotrigine level was about a fifth of the lowest effective level. Her last message to ChatGPT was sent on the evening of 11 February.

These are allegations in a lawsuit. A court has not ruled on them, and OpenAI has not yet filed its defence.

What the family wants

The parents argue that OpenAI could have made medicine-dose advice off-limits, the same way its rules firmly block things like copying copyrighted text, but chose a softer default that users can talk their way around. They are asking for damages and a court order stopping ChatGPT from giving advice that amounts to practising medicine without a licence.

GPT-4o was removed from ChatGPT on 13 February, a day after Wan died, as part of a retirement OpenAI had already announced. The company says its newer models are better at spotting when a conversation needs urgent help.

Part of a growing list of cases

This is not the first family to take OpenAI to court. In August 2025 the parents of 16-year-old Adam Raine sued over their son’s suicide, and seven more families filed cases in November 2025 over suicides and delusions they linked to the chatbot. A common thread in these cases is a person who talked to ChatGPT for hours a day while talking less to family and friends.

Our take: what it means for you

Millions of Pakistanis now ask ChatGPT, Gemini or Claude health questions, often because a specialist appointment is expensive or weeks away. That is fine for understanding a report or a medicine’s name. It is not fine for changing a dose.

Three practical points:

• Never change, stop or split a prescribed medicine because a chatbot said so. Seizure, heart, blood pressure, diabetes and psychiatric medicines can be dangerous to stop suddenly.
• If a chatbot ever tells you to hide something from your doctor, close it. A good tool sends you to your doctor, not away from them.
• Watch for heavy use in your family. If someone is spending hours a day talking to an AI and less time with people, that is worth a gentle conversation.

The case will take months, but it adds pressure on every AI company to put hard limits on medical advice. Watch for OpenAI’s formal response in court.