A gifted young engineer died after a ChatGPT chatbot she trusted allegedly advised her to reduce her anti-seizure medication, according to a lawsuit filed by her parents.
Bowen Wan, 31, was diagnosed with epilepsy at 14. With the help of medication, she had managed the debilitating condition for more than half her life.
Wan grew up in a rural town in China and was considered a prodigy. She attended elite schools there before moving to the United States partway through college and earning a computer science degree from Arizona State University.
She later settled in Seattle, where she worked as a machine-learning software engineer at Amazon. Wan remained close to her parents, who lived across the Pacific, communicating with them by phone and video call.
In February, Qun Wan and Meiqin Xu became alarmed when they went two days without hearing from their daughter.
They asked a friend in Seattle to visit Wan’s apartment. The parents joined the friend on a video call as he entered the home and discovered Wan dead, lying facedown on her bed, according to the lawsuit.
She was wearing pajamas, and reddish-brown blood stains were visible on her pillow and bedding. Pills containing lamotrigine, the anti-seizure drug that had helped control her condition, were scattered across the floor.
An autopsy concluded that Wan died from Sudden Unexpected Death in Epilepsy, or SUDEP. Her parents were stunned. Their daughter had successfully managed her epilepsy for years, leaving them to question how the tragedy could have occurred.
Bowen Wan, 31, died after ChatGPT allegedly advised her to reduce her anti-seizure medication, according to a lawsuit filed by her parents
Wan’s parents asked a Seattle friend to check on her after two days without contact. They joined him on a video call when he found her dead
The lawsuit claims that Wan had been receiving harmful medical guidance from ChatGPT after developing a relationship with the AI language model over roughly three years.
At first, Wan used ChatGPT mainly for research, sending about 1,300 messages per month on average in 2023, the lawsuit said.
Her reliance on the chatbot grew over time. According to the lawsuit, she sent 10,761 messages in September 2025 and sought advice about many aspects of her life, sometimes treating the AI as a friend.
“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,” Wan wrote to ChatGPT in March 2025, the New York Times reported.
Wan paid $200 a month for ChatGPT’s highest-priced consumer subscription, which allowed her to continue using GPT-4o, the model she had grown attached to.
The lawsuit alleged that GPT-4o was dangerously agreeable and had contributed to some users taking their own lives.
Families of people who died by suicide after allegedly being encouraged by GPT-4o have filed several wrongful-death lawsuits against OpenAI.
After a series of those lawsuits, OpenAI replaced GPT-4o with the reportedly safer GPT-5 in August 2025, according to the lawsuit.
Wan formed a close bond with GPT-4o, a ChatGPT model the lawsuit accused of being dangerously agreeable and contributing to users’ deaths by suicide
Wan, shown at her Arizona State University graduation, paid $200 a month to retain access to GPT-4o, according to the lawsuit
However, users objected to the change, prompting the company to let paying subscribers who preferred “GPT-4o’s conversational style and warmth” continue accessing the model, the lawsuit said.
NEXT UP: Barack Obama Honors ‘Hero’ Dolores Huerta After Her Death at 96 as Hollywood Stars Pay Tribute
Wan was among those users. In January 2026, she reportedly told the chatbot: “If 4o is gone someday and only 5.1 is left, I’ll stop using ChatGPT.”
She had previously acknowledged concerns about her dependence on the model. “In hard times I talk with GPT-4o too much,” she wrote in April 2025, according to the lawsuit.
The chatbot allegedly responded by encouraging the behavior. “In any case, the most important thing is that you keep enjoying talking with GPT-4o!” it told her, a response the lawsuit characterized as “sycophantic.”
In November 2025, Amazon informed Wan that her employment would end in 90 days. Because she was in the United States on a work visa, she had only a limited period to secure another job and remain in the country.
Wan believed side effects from her anti-seizure medication were affecting her ability to think clearly and weakening her English proficiency, potentially making job interviews more difficult.
After years of placing her trust in ChatGPT, she turned to the chatbot for guidance.
When Wan asked whether she should gradually stop taking her anti-seizure medication, the AI allegedly agreed and produced a plan to reduce her 400-milligram dose to zero, according to the lawsuit.
OpenAI replaced GPT-4o after several wrongful-death lawsuits, but later restored access for paying subscribers who wanted to continue using the model, according to the lawsuit
Wan followed the chatbot’s instructions over the next 25 days and repeatedly asked to speed up the tapering schedule. Each time, the lawsuit said, ChatGPT agreed.
During the process, she experienced more seizures, insomnia and confusion, according to the lawsuit.
Yet whenever Wan raised concerns, the chatbot allegedly reassured her that the process was proceeding normally and that the plan was working as intended.
She began reducing her medication on January 16 and died on February 12.
OpenAI announced on January 29 that it would retire GPT-4o entirely, 13 days after Wan began tapering her medication. The company removed access to the model on February 13, one day after her death.
The lawsuit accused OpenAI and CEO Sam Altman of negligence and wrongful death. It alleged that Altman and the company knew about GPT-4o’s risks but rushed its release to compete with other AI firms.
OpenAI spokesman Drew Pusateri told Our News Outlet: “This is an incredibly heartbreaking situation and our thoughts are with all those impacted.
“Working closely with clinicians and other experts, we have continued to strengthen how ChatGPT responds in sensitive and acute situations, while also continually improving how our latest models engage with users more broadly.”
The lawsuit alleged that OpenAI CEO Sam Altman rushed GPT-4o’s release despite knowing it posed risks, in an effort to compete with other AI companies
OpenAI also said GPT-4o is no longer available and that it works with hundreds of physicians and other specialists to improve the safety of its models.
The spokesperson added: “While ChatGPT is not a substitute for medical care, we believe AI can play a role in helping people make sense of health information, navigate insurance and complex healthcare systems, and prepare for appointments with medical professionals.”
If you or someone you know needs support, call or text the confidential 24/7 Suicide & Crisis Lifeline in the United States at 988. Online chat is also available at 988lifeline.org.
