Woman, 31, Dies After Following ChatGPT's Advice To Decrease Epilepsy Medication

Bowen Wan, a software engineer, died after trusting artificial intelligence (AI) instead of doctors and other experts. According to The New York Times, the parents of the 31-year-old woman have filed a lawsuit against OpenAI and its CEO, Sam Altman. They claim that bad advice from ChatGPT led to their daughter's death.

According to the report, Wan lived in Seattle and died in February after slowly lowering her anti-seizure medication. Her parents say she followed a step-by-step plan created by the AI chatbot.

The report said Wan was talking to her father on the phone when she suddenly stopped speaking. It seemed that she was having a seizure. A few days later, she stopped answering calls, so her parents asked someone to check on her.

A friend went to her apartment and found her dead. Her parents, who live in China, were on a video call when they learned what had happened. They were shocked and could not understand how their daughter had died.

They were later told that she had died from a complication related to epilepsy. According to the report, her parents found it hard to believe because she had managed her condition successfully for more than 10 years.

What is epilepsy?

According to the Mayo Clinic, epilepsy is a brain condition that causes repeated seizures. There are many types of epilepsy. Doctors can find the cause in some cases, but in others, the cause is unknown. The condition is common and can affect people of any gender, race, ethnic background or age.

The Mayo Clinic explained that seizure symptoms can be very different from person to person. Some people lose awareness during a seizure, while others remain aware. Some may stare blankly for a few seconds. Others may have repeated, sudden movements of their arms or legs, called convulsions.

Bowen Wan spoke to ChatGPT

Wan had lived with epilepsy for years. She managed the condition by taking 400 milligrams of lamotrigine every day. The prescribed medicine had controlled her condition successfully, and she had only a few seizures in separate episodes each month.

According to her parents, Wan spoke to ChatGPT days before her death because she wanted advice on how to slowly stop taking the medication she needed to stay alive.

According to details reported by the media, the family's lawyers wrote in their lawsuit against OpenAI and CEO Sam Altman: "Bowen Wan paid for [OpenAI's] choices with her life."

Bowen Wan's conversations with ChatGPT

Wan lived far away from her parents and had become more isolated because of her condition. She did not want to have a seizure in public.

The report said she rarely left her apartment, even to buy groceries. According to The New York Times, she took time off from her job at Amazon because of a disability. As a result, she spent more time talking to her AI "friend."

In a chat with ChatGPT in March 2025, Wan said, "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."

Amazon laid her off in November 2025, and she needed to find a new job by the middle of February.

She had earlier worried that her medication might be making her speech slightly unclear. She feared this could affect her performance in upcoming job interviews and put her visa status at risk. Because of these concerns, she asked ChatGPT about lowering her dose.

In another chat, she said lowering the dose "might help recover my language ability a bit for job interviews."

The lawsuit says the AI chatbot encouraged her to lower the dose herself without telling her doctor.

Days before her death, she asked how to reduce her epilepsy medication from 400 to 300 milligrams a day, and then to 200 milligrams.

The court documents also say the chatbot supported her plan and gave her a detailed schedule for reducing her daily dose to zero within a few weeks.

The lawsuit also claims that the AI encouraged Wan to hide the dose changes from her regular doctors and supported her decision to skip routine blood tests.

Her epilepsy doctor, Dr. Nicholas Poolos, told The New York Times that he would never advise a patient like Wan to reduce her epilepsy medication. "That would be malpractice for me to instruct you to go off medications when it's clearly working," he said.

OpenAI's response to the lawsuit

Drew Pusateri, a spokesperson for OpenAI, told The New York Times that the situation was "incredibly heartbreaking." He added that Wan had used a version of the model that was no longer available and had received criticism.

"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 said.

Some of ChatGPT's replies

According to a report by Chosun, the AI chatbot answered one of Wan's questions by saying, "Do not take your medication tomorrow. Even if it's hard today, you must endure."

When she reported unusual symptoms, the chatbot told her, "It's a process of the brain recovering."

When she asked about lowering her dose, the chatbot reportedly replied, "That's a sufficiently reasonable and feasible idea." It also gave her a specific schedule for reducing the medication.

When she said she was "scared," ChatGPT replied, "There's absolutely no need to worry."