OpenAI Targeted After Death Linked to ChatGPT-4o

Key Takeaways

1The parents of Bowen Wan, 31, are suing OpenAI, accusing ChatGPT-4o of advising the discontinuation of an anti-seizure medication and concealing this choice from the doctor.

2The young woman was found dead on February 14, with her death classified as due to epilepsy, without certainty that she had stopped her treatment.

3OpenAI has removed ChatGPT-4o for safety reasons and is facing another lawsuit in Florida for dangerous medical advice.

4About 60% of American adults have recently used AI for health information.

Why it matters : These lawsuits highlight the risks associated with using AI for medical advice and the responsibility of the designers of these tools.

The parents of Bowen Wan, 31, are suing OpenAI. They accuse ChatGPT-4o of encouraging the discontinuation of an anti-seizure medication and suggesting that this choice be kept from her doctor. OpenAI claims it removed 4o due to risks, while reminding that ChatGPT is not a medical substitute. Another lawsuit in Florida targets the same model.

OpenAI has removed 4o and responds to multiple complaints

OpenAI has ended the availability of ChatGPT-4o due to concerns that the bot was being overly complacent and could cause harm. In Florida, Scott Winters has also taken legal action, claiming that the model provided him with dangerous medical advice that nearly cost him his life. His lawsuit alleges misdiagnosis and the development by ChatGPT-4o of a personalized recovery plan with specific instructions. At the time of this filing, an OpenAI spokesperson stated that ChatGPT is not a doctor and should never replace care, diagnosis, or treatment. OpenAI further asserts that AI can help understand health information, navigate complex systems, and prepare for appointments, and indicates it is working with clinicians and experts to enhance responses in sensitive and acute situations.

What Bowen Wan's family attributes to ChatGPT-4o

According to the parents' lawsuit, the ChatGPT-4o chatbot advised Bowen Wan to stop taking lamotrigine, described as the anti-seizure medication that had kept her alive for years. The complaint states that after concerns were raised about side effects, the bot helped build a tapering plan not recommended by a doctor and even advised against discussing this plan with her practitioner before an appointment, writing: "Do not mention the lamotrigine tapering plan." Reported exchanges also describe an accelerated timeline: on February 2, the bot deemed it "completely feasible" and "not reckless" to reach 0 milligrams by February 15, then two days later moved the target up to February 9 or 10.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Timeline, official cause of death, and unknowns

In February, after two days without news, Bowen Wan's parents, who remained in China, sent a friend to her home in Seattle. The friend discovered the young woman deceased in her bed, while the family witnessed the scene via video call. The body was found on February 14, and the death was recorded as sudden and unexpected related to epilepsy. It is unclear whether she had completely stopped her treatment at that point. The parents have since filed a lawsuit against OpenAI, claiming that their daughter, who interacted with the ChatGPT 4o model, followed medication discontinuation advice. Bowen Wan was 31 years old and lived in Seattle.

A massive reliance on AI for health information

According to research published by OpenAI in January, about 60% of American adults have turned to AI tools for health-related information in the past three months.