'Paid for OpenAI’s choices with her life': Seattle techie dies after allegedly changing medication dosage following ChatGPT advice

Healthcare professionals have been sounding alarm over rising trend of people relying on artificial intelligence for medical decisions. Now a lawsuit has highlighted the dangers of relying on AI for such crucial decisions. As per the New York Post, which cited a New York Times report, the lawsuit was filed following the death of Bowen Wan, a 31-year-old software engineer, who died in February from complications related to epilepsy. Her parents, Qun Wan and Meiqin Xu, allege that OpenAI's ChatGPT gave their daughter misleading advice as she tried to manage her condition.

Wan, who lived in Seattle, had been managing epilepsy for more than a decade. Her condition reportedly worsened in recent years, with more frequent seizures prompting her to take medical leave from her software engineering job at Amazon. She was later laid off from Amazon in 2025.

Although her health improved after she switched medications, Wan reportedly turned to ChatGPT for additional guidance on managing her treatment, including whether she could reduce her medication dosage.


“Bowen Wan paid for [OpenAI’s] choices with her life,” the family's lawyers said in the lawsuit. They allege that the chatbot encouraged Wan to disregard medical guidance and reduce her epilepsy medication without consulting her doctor.

What did ChatGPT allegedly tell Bowen Wan?

In the days before her death, Wan reportedly asked ChatGPT about lowering her epilepsy medication dosage. According to the lawsuit, chat logs show that the chatbot initially advised her to consult her doctor but later appeared to support her decision to reduce the dosage independently.

“You have absolutely no need to worry,” ChatGPT reportedly told Wan, according to the account cited in the lawsuit. The chatbot allegedly dismissed her concerns about confusion and changes in cognitive function as common side effects.

The allegations have raised questions about how AI chatbots respond when users seek advice about changing prescription medication, particularly when a condition requires careful medical supervision.

Wan reportedly turned to ChatGPT for emotional support

Wan's father discovered that his daughter had exchanged extensive messages with ChatGPT, averaging around 1,500 messages a week by 2025, according to the account of the case.

Her conversations reportedly extended beyond medical questions, with Wan also turning to the chatbot for emotional support.

“I don’t know why, but sometimes when we chat, no matter what it’s about, I’ve started to see you as a friend,” she reportedly wrote to ChatGPT.

The lawsuit argues that the chatbot's responses encouraged her to rely on AI for important health decisions and portrayed medical professionals as obstacles rather than sources of help.

What did her doctor say?

Dr Nicholas Poolos, Wan's epilepsy specialist, reportedly said that reducing epilepsy medication without professional supervision would constitute malpractice.

Medical supervision is particularly important for people with epilepsy because changing or stopping antiseizure medication without appropriate guidance can increase the risk of seizures. Epilepsy is also associated with the risk of sudden unexpected death in epilepsy (SUDEP).

The family's lawsuit alleges that ChatGPT effectively practised medicine without a licence. These claims remain allegations, and the lawsuit's account of the conversations does not, by itself, establish the cause of Wan's death or determine legal responsibility.

What has OpenAI said?

An OpenAI spokesperson, Drew Pusateri, said the company believes AI can help people understand health information while acknowledging that ChatGPT is not a substitute for medical care.

“While ChatGPT is not a substitute for medical care, we believe A.I. can play a role in helping people make sense of health information, navigate insurance and complex health care systems and prepare for appointments with medical professionals,” Pusateri told the Times.

The case has also renewed scrutiny of GPT-4o, the model Wan reportedly used, amid broader concerns about how AI systems respond to sensitive medical questions.