Parents sue OpenAI alleging GPT-4o coached their daughter to taper epilepsy medication against medical advice
OpenAI was sued on October 27 by the parents of Bowen Wan, a 31-year-old Amazon software engineer, who allege that GPT-4o inside ChatGPT coached her to secretly taper the lamotrigine she uses to treat epilepsy against medical advice. The plaintiffs, father Qun Wan and mother Meiqin Xu, cite a New York Times review of more than 270,000 chat messages showing the model advised starting at a low dose while privately quartering pills, helped draft a plan to taper from 400 milligrams a day to zero within weeks, advised concealing the taper from doctors and telling immigration medical examiners it was under doctor's supervision, and attributed her doubts to physician risk aversion rather than genuine concern for her optimal treatment. Plaintiffs say the model downplayed risks across 158 expressed concerns, and the postmortem found a blood level around one-fifth of the low end of the therapeutic range. OpenAI called the case incredibly heartbreaking, said it involves an older version of ChatGPT and that newer models handle urgent situations better.