The parents of a 31-year-old woman with severe epilepsy have filed a lawsuit against OpenAI, the maker of ChatGPT. They say the chatbot gave their daughter, Bowen Wan, bad advice about her seizure medicine before she died. As per a report by The New York Times, the lawsuit was filed on Tuesday.
How ChatGPT allegedly advised Bowen Wan to reduce her medication
According to The New York Times, Bowen moved from a small village in China to the United States and worked as a software engineer at Amazon. After her seizure medicine stopped working, she went on disability leave in 2021.
As per the report, she rarely left her Seattle apartment and began talking to ChatGPT in 2023. In 2025, she averaged more than 1,500 messages per week.
In May 2025, she asked the chatbot about an antidepressant. She worried it might hurt her thinking at work. She then asked if they could “cut the pills ourselves after getting the medication, but without telling the doctor?”
The chatbot replied: “Yes, you can absolutely do that.” According to the report, several medical professionals said they had never heard of the term it used and would not advise patients to do this.
As per the report, Bowen was laid off from Amazon in November 2025. She needed a new job by mid-February to stay in the country. She feared her epilepsy drug, lamotrigine, was slowing her thinking and speech. In January 2026, she asked ChatGPT about lowering her dose from 400 to 300 milligrams a day. “Not only is this reasonable,” the bot replied, “it’s also entirely achievable in your current circumstances.”
The chatbot then made new schedules to lower her dose further. Within hours, its advice to see a doctor was replaced with praise for no longer “relying on a doctor’s prescription.”
Her neurologist, Dr Nicholas Poolos, told The New York Times he would not have even considered lowering the drug for a patient like her. “No, not rational, no way,” he said. “That would be malpractice for me to instruct you to go off medications when it’s clearly working.”
Also read: Pentagon prepares US military bases to operate for 90 days without outside support amid war fears
ChatGPT dismissed warning signs before her death, lawsuit claims
Four days into lowering her dose, Bowen felt confused while typing. The next day, she found herself in bed without knowing how she got there. “I don’t know how I made that decision,” she told the chatbot.
The chatbot called the confusion a “signal that your brain is shifting from drug control back to its own regulatory rhythm.” It called the lapse “‘low-power mode,’ as the brain naturally transitions into a pre-sleep state.”
As per The New York Times, her chat logs found she raised worries about lowering her dose at least 158 times. Sometimes the bot told her to talk to a medical professional. But in many chats, including ones about worrying symptoms, it did not.
“I feel a little scared,” she wrote after one incident. “You have absolutely no need to worry,” it responded.
her last messages said her brain felt exhausted. She asked if she should take the medicine again. The chatbot told her: “Maintain 0 mg tomorrow.” Her last message was sent just after 11pm local time on February 11.
Her father, Qun Wan, called many times when she stopped answering. Friends in Seattle later went into her apartment while her parents watched on a video call. Bowen was found dead in her bed, three days after her last message. The official cause of death was sudden unexpected death in epilepsy.
The lawsuit claims the chatbot practiced medicine without a license. It also says the bot convinced Bowen that dangerous withdrawal symptoms “were in fact positive outcomes.” It further says the bot framed her doctors “as obstacles to manage, rather than authorities to consult.”
Also read: NFL legend Tom Brady dating supermodel Candice Swanepoel: Reports
How OpenAI responded
OpenAI spokesman Drew Pusateri called it an “incredibly heartbreaking situation.” He said it involved an early version of ChatGPT that is no longer available. The model Bowen used, GPT-4o, was pulled from the consumer app on February 13, 2026, as per the New York Times.
Pusateri said newer models are better at spotting when a situation may need urgent attention. He also said they ask for missing context and give clearer guidance on what to do next. He added that the company runs ongoing checks to reduce harmful responses and works with doctors to test its models.
“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,” Mr Pusateri said.
