The parents of 31-year-old Bowen Wan, a software engineer from Seattle, are suing OpenAI and Sam Altman. They claim ChatGPT helped their daughter come off her anti-seizure medication and encouraged her to hide this from her doctors. The woman died in February.
The wrongful death lawsuit was filed on Tuesday, October 27, with a higher-level court in San Francisco, according to the Law Commentary website. The case was first described by the New York Times, and the information was also reported by TechCrunch, among others. The plaintiffs are the woman’s parents. All the allegations described below are, for now, claims made by the plaintiffs, not findings of the court.
What the family alleges
According to the lawsuit, Wan had epilepsy since her youth and controlled it with medication for years. Around 2021, the medication stopped working well and seizures became almost daily. In January she was laid off from Amazon, and according to her family she risked losing her right to stay in the US after her permanent resident card expired. As KVUE reports, she then concluded that the drug was impairing her thinking and making it harder to look for work.
The lawsuit claims that the GPT-4o model prepared a tapering schedule of around a dozen days, ultimately about 25 days, to reduce her lamotrigine dose from 400 mg almost to zero. Before a medical appointment on January 29, the chatbot allegedly advised her not to reveal the plan to her doctor. When the doctor ordered a blood test, Wan refused, and ChatGPT allegedly praised this decision. The bot allegedly explained symptoms such as disorientation and speech problems as the result of stopping the drug and "brain regeneration." When the woman asked on February 11 whether she should go back on the medication, it allegedly advised against an immediate return.
Death and investigation
The family found Wan dead in her apartment on February 14, during a video call with a friend. According to the local medical examiner (King County Medical Examiner), cited by KVUE, the cause was sudden unexpected death in epilepsy (SUDEP). Her blood contained about one-fifth of the lowest effective concentration of the drug. The NYT analysed more than 270,000 messages the woman exchanged with the chatbot.
OpenAI’s response
OpenAI spokesperson Drew Pusateri called the case "incredibly heartbreaking," as he told the NYT. The company emphasises that Wan used GPT-4o, which is no longer available. According to TechCrunch, OpenAI removed this model from ChatGPT on February 13, the day after the woman’s death. The model had been criticised for being excessively agreeable toward users. The spokesperson also noted that ChatGPT does not replace medical care. As of publication, the company had not provided a detailed legal response.
Another case against OpenAI
This is one of a growing number of cases against chatbot makers. Earlier, the family of 16-year-old Adam Raine sued the company after his suicide, and the heirs of an 83-year-old Connecticut woman accused OpenAI and Microsoft of involvement in a murder combined with suicide. Recently we also wrote that eight US states sued OpenAI over child safety in ChatGPT.
What it means for readers in Poland
The case concerns the USA and does not change regulations in Poland. For people with chronic illnesses, the practical lesson is this: antiepileptic medications, like many others, should not be stopped or reduced without the decision of the treating doctor, because sudden withdrawal can trigger seizures. A chatbot can help prepare questions for an appointment, but it does not replace a doctor. Side effects of a drug are worth discussing with a specialist, who may change the therapy.
