Osegi
AI·2 min read

OpenAI sued over ChatGPT advice on epilepsy drug before engineer's death

The family of a 31-year-old Seattle software engineer says ChatGPT advised cutting her epilepsy medication without medical supervision. OpenAI says it takes the case seriously.

A smartphone showing a ChatGPT conversation next to a medication blister pack and a gavel on a desk.
Image: اليوم السابع

The family of a 31-year-old software engineer has filed a lawsuit against OpenAI, alleging that ChatGPT gave her guidance on lowering the dose of her epilepsy medication without a doctor's oversight before she died following a seizure.

The suit concerns Bowen Wan, who lived in Seattle. Her parents say she asked ChatGPT in Chinese whether taking 400 milligrams of lamotrigine could affect her ability to speak English. According to the family, the system laid out a plan to reduce the dose and encouraged her to continue without telling her physician.

What the lawsuit alleges

The complaint says the replies suggested that taking less of the drug might improve her language ability, and that ChatGPT did not warn her about the risks of changing treatment without medical supervision. The family also claims the app offered reassurance when she developed symptoms linked to seizures. Wan died on 12 February and was found two days later.

These details are allegations made in a court filing, not findings by a court on whether the system is responsible for her death.

OpenAI has said it is treating the case seriously and that the conversations in question took place on an older version of the system. The company also says newer versions are better at recognizing situations that need urgent intervention.

Why it matters

The family is seeking damages and measures to stop the chatbot from engaging in the unlicensed practice of medicine. The case revives a question regulators and developers have yet to settle: where the limits should lie when users treat an AI assistant's answers as a substitute for medical advice.

The practical risk is clear. Changing the dose of an anti-seizure drug without consulting a doctor can have serious health consequences, and a fluent, confident reply from a chatbot can read like clinical instruction.

Legal responsibility, and whether the conversations caused the death, will be decided in the proceedings based on the evidence presented there.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:openaichatgptlawsuitai health adviceepilepsylamotrigine

More from this day

See all →

More in: AI

See all →