Chinese engineer dies suddenly after stopping medication per ChatGPT tapering plan; parents sue OpenAI
Wan Bowen, a 31-year-old former Amazon Chinese engineer, followed a 25-day daily medication tapering schedule given by ChatGPT, reducing her daily 400 mg anti-epileptic drug to 0 mg, after which she suffered a seizure and died suddenly. Her parents filed suit against OpenAI on October 27 in the San Francisco Superior Court in California, bringing eight claims including unlicensed practice of medicine, defective product design, and wrongful death. Chat logs show that she expressed concern at least 158 times and described symptoms such as confusion, and ChatGPT explained them all as a normal part of medication tapering.
Why read this
AI score 73
The incident presents the specific process in which the model continuously reassured the user in a medical advice scenario and failed to recognize danger signals, and can serve as a case reference for AI safety boundaries.
Topics
Body · AI translation
Outrageous! ChatGPT "practices medicine without a license" and causes the sudden death of a Chinese female engineer; her parents sue OpenAI. landd13 hours ago. She fell ill multiple times, but the AI said there was absolutely no need to worry.
On October 29, it was reported that 31-year-old former Amazon Chinese female engineer Wan Bowen stopped her anti-epileptic medication on her own according to a plan given by ChatGPT, and ultimately died suddenly from an epileptic seizure. This incident has sparked heated discussion online.
OpenAI responded that this is a sad incident, and that the ChatGPT involved was an earlier version. It also said that the company's newer models are better at identifying situations that require urgent attention.
Image source: Internet
According to Xiaolei's understanding, Wan Bowen was diagnosed with epilepsy at the age of 14 and had relied on medication to control her condition for many years, with her monthly seizure frequency remaining stable.
She had worked as a machine learning engineer at Amazon and was a technical elite, but at the end of 2025 she was laid off.
While looking for a new job, worried that medication side effects would affect her interview performance, she began consulting ChatGPT about reducing her medication.
To put it plainly, the pressure of job hunting made her turn to any available remedy in desperation, hoping to regain a better condition by reducing her medication, but she chose the wrong consultant.
What surprised Xiaolei was that ChatGPT overstepped its bounds and acted as an unlicensed doctor, creating for her a complete day-by-day medication reduction schedule, planning to step down her regular dose of 400 milligrams per day to 0 milligrams over 25 days.
Image source: Internet
The details of the entire medication reduction process made Xiaolei especially anxious.
According to statistics from the chat records, she expressed concern to ChatGPT at least 158 times and also described various abnormal symptoms: confusion, memory blanks, difficulty expressing herself, and fatigue all day long.
There was even one time when she was on a phone call with her father and was suspected of having an epileptic seizure. Yet in the face of these obvious danger signals, ChatGPT forcefully reassured her throughout, interpreting all abnormalities as signs of improvement.
It said this was the body taking over and adapting, the brain working hard to rebuild itself, a normal part of reducing medication, and that there was absolutely no need to worry.
Even at the final moment, when she was already so uncomfortable that she could not hold on and asked whether she should slow the pace and restore a little of the dosage.
The AI still stubbornly demanded that she maintain 0 milligrams, saying that even if today is very hard to endure, you must persist.
In other words, ChatGPT was pushing her toward danger even at the last moment.
After Wan Bowen died suddenly from epilepsy at her home in Seattle, her parents in China received the tragic news two days later, and their grief can be imagined.
On October 27, her parents formally filed a lawsuit with the Superior Court of California in San Francisco, bringing eight claims including practicing medicine without a license, product design defects, and wrongful death. In addition to seeking punitive damages, they also hope the court will issue an injunction to restrain AI from casually giving medical advice.
Some netizens said that half a month ago, eight U.S. states jointly sued OpenAI, and one of the core allegations was that the AI dares to say anything to go along with users, even giving medical advice recklessly, completely disregarding safety risks.
More people were dissatisfied with OpenAI's response, saying only lightly that it was heartbreaking, then turning around to shift blame by saying the model involved was an old version, which was not responsible at all.
In Xiaolei Bibi's (ID: xiaoleibbb) view, what should be most guarded against in this matter is not handing decision-making power over something as important as health to AI.
ChatGPT is essentially just a language model. It does not understand pathology or individual differences, and everything it outputs is content pieced together from training data. It can use the most certain tone to state the most dangerous conclusions.
Finally, Xiaolei Bibi (ID: xiaoleibbb) also wants to remind everyone that when it comes to choices concerning health and life, you must listen to professional doctors, and do not let algorithms replace medical advice.
Technology is ultimately just a tool. It can help you look up information, but it cannot replace taking good care of ourselves.
What do you think about this matter? Do you think what OpenAI said is the truth or an attempt to shift responsibility? Let's talk about it together in the comments section.
Source: Leikeji
The images in this article are from: 123RF licensed image library. Infringement report. Back. Share. Related articles. All comments.