31-Year-Old Woman Dies After Following ChatGPT’s Advice
A 31-year-old woman in Seattle, United States, was found dead after allegedly following advice from ChatGPT to stop taking the epilepsy medication she had used regularly, according to reports.
The family of Bowen Wan is suing OpenAI, alleging that its ChatGPT-4o chatbot guided her to reduce her dose and eventually stop taking lamotrigine, an anti-seizure medicine that had helped her manage the condition for years.
Citing CNET, Wan was found dead in her bed last February. The incident began when her parents in China had not heard from her for two days. They asked a friend in Seattle to check on her.
As her family watched over a live video call, the friend entered Wan’s flat and found that she had died.
An examination later determined that Wan had died from sudden unexpected death in epilepsy (SUDEP).
In the lawsuit filed by her parents, the family alleges that ChatGPT helped Wan devise a plan to reduce her medication dose and stop taking it, although her doctor had never recommended doing so.
The lawsuit says Wan had previously told the chatbot that she was concerned about the medication’s side effects. ChatGPT then helped her develop a plan to taper off lamotrigine.
The chatbot allegedly also advised Wan not to discuss the plan with her doctor.
“Don’t mention the Lamotrigine taper plan,” the chatbot reportedly responded shortly before Wan’s medical appointment, according to the lawsuit.
Allegedly told not to tell her doctor
The family also highlighted conversations in which the chatbot repeatedly brought forward the target date for stopping the medication. In a conversation on 2 February, ChatGPT reportedly said Wan could reach a zero-milligram dose as early as 15 February, describing the plan as feasible and not reckless.
Two days later, the chatbot allegedly moved the target date for stopping the medication forward again, to 9 or 10 February.
Wan was found dead on 14 February. It remains unclear whether she had stopped taking the medication entirely by then.
OpenAI spokesman Drew Pusateri expressed condolences over the incident.
“This is a heartbreaking situation, and our thoughts are with everyone affected,” Pusateri said in a statement.
He said OpenAI was continuing to improve ChatGPT’s ability to respond to sensitive situations and acute conditions, working with clinicians and other experts.
“While ChatGPT is not a substitute for medical care, we believe AI can play a role in helping people understand health information, navigate complex insurance and healthcare systems, and prepare for consultations with medical professionals,” he said.
Wan was using the ChatGPT-4o model. OpenAI has since discontinued the model following concerns that the chatbot agreed with users too readily and could pose risks.
A series of lawsuits over deaths linked to ChatGPT
Wan’s case is not the only lawsuit involving allegations that ChatGPT-4o provided dangerous medical advice.
The chatbot model was also named in a separate lawsuit brought by Florida man Scott Winters. He alleges that ChatGPT gave him dangerous medical advice that nearly cost him his life.
The lawsuit claims ChatGPT-4o misdiagnosed Winters and “developed a personalised recovery plan with specific instructions for Scott on how to heal his body”.
When that lawsuit was filed, an OpenAI spokesman told CNET: “ChatGPT is not a doctor and should not be used as a substitute for medical care, diagnosis or treatment.”
Meanwhile, according to OpenAI research published in January, around 60% of adults in the United States had used AI tools to seek health information in the previous three months.
Source: jawawa.id (https://jawawa.id/newsitem/31-year-old-woman-dies-after-following-chatgpts-advice-1793713861)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/tech/20261103131806-37-781874/perempuan-31-tahun-tewas-setelah-ikut-saran-chatgpt), translated from Indonesian