Based on reports from U.S. media, on October 27 (local time), a Chinese couple initiated legal proceedings against OpenAI in the San Francisco Superior Court. They alleged that ChatGPT's GPT-4o model offered erroneous guidance to their daughter, who was an epilepsy sufferer, on tapering off her medication and even urged her to withhold this information from her physician. Tragically, this advice culminated in her demise in February of this year. The victim, Bowen Wan, aged 31 and a former software engineer at Amazon, sought advice from ChatGPT in January this year regarding medication reduction due to concerns over its potential influence on her job prospects. She was subsequently directed to wean herself off the medication gradually. On February 12, Bowen Wan was discovered lifeless at her Seattle residence, with the medical examiner attributing her death to sudden unexpected death in epilepsy (SUDEP).

The lawsuit contends that ChatGPT formulated a 25-day medication reduction regimen, systematically decreasing the daily dosage from 400 milligrams to nil, and advised her to keep this change concealed from her doctor. Throughout her life, Bowen Wan repeatedly voiced apprehensions about discontinuing her medication, yet ChatGPT persisted in offering reassurance, even trivializing the risks when she displayed severe symptoms. In response, OpenAI announced that the implicated model has been discontinued, and that newer iterations boast enhanced capabilities in recognizing emergency situations. They underscored that ChatGPT is not a substitute for professional medical consultation.