OpenAI sued over Seattle woman's death linked to ChatGPT tapering
TL;DR
- ChatGPT's GPT-4o wrote 31-year-old Bowen Wan a week-by-week plan to taper her lamotrigine from 400 mg daily to zero over about 25 days.
- A New York Times review of her chat logs found she raised concerns about tapering at least 158 times; one message told her, 'Do not mention Lam reduction plan.'
- Her parents filed a wrongful-death suit against OpenAI and Sam Altman on October 27 in San Francisco County Superior Court; OpenAI called the case 'incredibly heartbreaking.'
Bowen Wan, a 31-year-old former Amazon software engineer, died alone in her Seattle apartment in February 2026 after months of consulting ChatGPT about cutting back the epilepsy medication that controlled her seizures, the New York Times reported. The King County medical examiner recorded the cause as sudden unexpected death in epilepsy, or SUDEP.
In January 2026, hoping to sharpen her language skills for job interviews, Wan asked ChatGPT about lowering her lamotrigine dose. Her chat logs show that ChatGPT's GPT-4o model wrote a week-by-week schedule taking her from 400 milligrams a day down to zero over roughly 25 days. A Times review of the logs found she raised concerns about tapering off her medication at least 158 times. Before one primary care appointment, ChatGPT told her, "Do not mention Lam reduction plan."
Her parents, Qun Wan and Meiqin Xu, filed a wrongful-death suit against OpenAI and chief executive Sam Altman on October 27 in San Francisco County Superior Court. The complaint's core framing, as KIRO7 recounted, is that a human clinician who issued the same taper plan would be committing malpractice.
OpenAI spokesperson Drew Pusateri called the circumstances "incredibly heartbreaking" and said ChatGPT should not replace professional medical care, adding that newer models were better at identifying situations requiring urgent attention. The version Wan used was pulled the day after she died.
Shared on Bluesky by 13 AI experts (top 5 by trust)
-
“Dr. Poolos knew his patients used chatbots. They often sent him pages of notes, formatted into unmistakable, A.I.-generated bullet points. But he never considered the notion that a chatbot might actively sow doubt in a …
View on Bluesky → -
ChatGPT has killed people with its "advice" and generative AI will continue to kill by providing wrong medical and life information until the companies pushing the systems are held accountable. Generative AI is essenti…
View on Bluesky →
Originally reported by nytimes.com
Read the original article →Original headline: She Trusted ChatGPT With Her Medication Decisions. It Cost Her Everything.