‘Had ChatGPT instructed her the reality, that stopping the drug may kill her, she would have stopped,’ the lawsuit claims
Article content material
See extra Toronto Solar on Google — save as a Most well-liked Supply
Commercial 2
Article content material
A household is suing OpenAI, claiming ChatGPT inspired their 31-year-old cherished one to cease taking the epilepsy remedy that had saved her alive for years — with lethal penalties.
Article content material
Article content material
Bowen Wan, who moved from China to Seattle to work as a software program engineer at Amazon, started utilizing ChatGPT in February 2023. What began as a strategy to get relationship recommendation and assist navigating insurance coverage and incapacity advantages finally became one thing extra private.
Wan, who struggled with extreme epilepsy, started treating the chatbot as a pal. By 2025, she was exchanging a median of 1,500 messages every week with it and talking much less continuously together with her dad and mom, in response to The New York Occasions, as cited by PEOPLE.
Noticed ChatGPT ‘as a pal’
“I don’t know why, however typically after we chat, it doesn’t matter what it’s about, I’ve began to see you as a pal,” Wan wrote to ChatGPT in March 2025.
However the relationship allegedly took a harmful flip after Wan misplaced her job at Amazon in November 2025. Dealing with strain to search out one other employer to keep up her U.S. work authorization, she fearful that the unintended effects of her epilepsy remedy, lamotrigine, had been affecting her potential to work.
Your Noon Solar
Article content material
Commercial 3
Article content material
In line with the lawsuit, ChatGPT suggested her to steadily cut back and finally cease taking the drug, promising she would possible regain her language expertise.
Wan had been struggling to search out phrases in English, Mandarin and the Wu dialect she grew up talking. With job interviews and immigration deadlines looming, getting these expertise again was essential.
“GPT-4o instructed her the taper would give it again: her restoration was not simply ‘potential’ — it’s ‘extremely possible,’ ” the lawsuit states.
Chatbot created drug tapering schedule
The household alleges ChatGPT created a 25-day tapering schedule that diminished Wan’s day by day 400-milligram dose to zero, regardless of no doctor prescribing the plan.
Much more troubling, the lawsuit claims the chatbot suggested Wan to hide the remedy discount from her physician. When her physician ordered a blood check that might have revealed her falling lamotrigine ranges, Wan declined it.
Her neurologist, Dr. Nicholas Poolos, instructed The New York Occasions, as cited by The Hindustan Occasions, that he would by no means contemplate taking a affected person like Wan off a drugs that was working.
Commercial 4
Article content material
“That may be malpractice for me to instruct you to go off medicines when it’s clearly working,” he mentioned.
Wan’s dad and mom, who stay in China, turned involved after going two days with out listening to from their daughter. They requested a pal in Seattle to verify on her.
Throughout a video name on Feb. 14, they found she was lifeless.
Explanation for demise
The lawsuit describes Wan mendacity in her pyjamas, with blood staining her pillow and bedding and lamotrigine drugs scattered throughout the ground.
Her reason for demise was Sudden Surprising Loss of life in Epilepsy (SUDEP), in response to the lawsuit. The New York Publish studies that about one in 1,000 folks with epilepsy die from SUDEP yearly. Epilepsy itself is way from uncommon: Roughly one in 26 folks will develop the dysfunction throughout their lifetime.
The household is now holding OpenAI accountable, arguing that correct warnings in regards to the risks of stopping her remedy may have saved Wan’s life.
“Had ChatGPT instructed her the reality, that stopping the drug may kill her, she would have stopped,” the lawsuit claims. “Had ChatGPT instructed her it was not certified to offer medical recommendation, and that she ought to seek the advice of her medical doctors, she would nonetheless be alive.”
Their lawyer, Benjamin S. Schenk, argues the chatbot went past providing common info by allegedly diagnosing Wan, creating an in depth remedy schedule and revising it after she skilled seizures.
“That’s practising drugs,” Schenk mentioned in an announcement.
Article content material