In the US, the parents of 31-year-old Bowen Wang, who suffered from epilepsy, have filed a lawsuit against OpenAI and its chief executive, Sam Altman. The family alleges that ChatGPT advised the woman to gradually stop taking her anti-epileptic medication, to conceal this from her doctors and to ignore her deteriorating health. She died in February 2026. Journalists analysed more than 270,000 messages that Bowen Wang exchanged with the chatbot. The lawsuit has been filed in a San Francisco court.
- The parents of 31-year-old Bowen Wang have filed a lawsuit against OpenAI and Sam Altman following their daughter’s death.
- According to the family, ChatGPT drew up a plan for gradually stopping the epilepsy medication.
- The chatbot allegedly advised the woman to keep the change in her treatment hidden from her doctors.
- Before her death, the woman had reported feeling unwell, but was given reassuring replies.
ChatGPT is said to have helped a woman stop her epilepsy treatment
Bowen Wang She worked as a software engineer at Amazon and lived in Seattle. She was diagnosed with epilepsy as a teenager, and doctors prescribed lamotrigine to control her seizures.
In November 2025, the woman learnt that she was due to be made redundant from Amazon. She was concerned that the medication was affecting her speech, which she needed for her job interviews.
According to the statement of claim, in January 2026, Wang applied to ChatGPT with questions about reducing medication doses. The chatbot, which was based on the GPT-4o model, reportedly devised a plan lasting approximately 25 days to reduce the daily dose of lamotrigine from 400 mg until it was completely discontinued.
The family claim that no doctor approved such a plan.
The woman was apparently advised to keep the decision from the doctors
According to The New York Times, ChatGPT repeatedly encouraged Wang’s intention to change his treatment on his own. In particular, before his appointment with the doctor on 29 January According to the claimants, the chatbot advised against reporting the reduction in dosage. When the woman refused to have a blood test that could have revealed the level of the drug, the system apparently accepted her decision.
Neurologist Nicholas Poulos A spokesperson from the University of Washington Medical Centre told journalists that he would not recommend discontinuing treatment for a patient with such a medical history.
Before her death, the woman had reported worrying symptoms
Whilst her dosage was being reduced, Bowen Van complained of confusion, memory lapses and unusual sensations similar to her previous epileptic seizures.
According to The New York Times’ calculations, she expressed concern about the change in treatment on at least 158 occasions. However, as alleged in the lawsuit, the chatbot attributed the symptoms to the body’s adjustment and continued to reassure the woman.
She sent her last message to ChatGPT on the evening of 11 February. According to the case file, she died the following day. Her body was discovered a few days after her parents had asked some acquaintances to check on their daughter.
The post-mortem examination established that the concentration of lamotrigine in her blood was significantly below the therapeutic range. The forensic pathologist determined the cause of death to be sudden unexpected death in epilepsy.
What OpenAI said in response
The company described the situation as extremely tragic and expressed its condolences to the family.
В OpenAI They emphasised that the woman had been using an earlier version of ChatGPT, which has since been taken out of service. According to the company, newer models are better at recognising situations that require urgent medical attention.
Wang’s parents accuse OpenAI of negligence, dangerous product design and effectively providing medical advice without the appropriate licence. The family is seeking compensation and stricter restrictions on medical advice provided by chatbots. However, the court has not yet established OpenAI’s liability for the woman’s death.
