Parents sue OpenAI, allege ChatGPT urged daughter to stop epilepsy medication
In Short
The lawsuit says that ChatGPT told an epilepsy patient to change their medicine. It also says that ChatGPT told the patient to keep the changes secret, from doctors. The lawsuit adds that ChatGPT ignored symptoms that could have been very serious.
Share :
Parents sue OpenAI, allege ChatGPT urged daughter to stop epilepsy medication
The parents of Bowen Wan, a 31-year- software engineer who worked at Amazon have filed a lawsuit against OpenAI. They say that ChatGPT was involved in decisions that led their daughter to stop taking her epilepsy medication without a doctor’s approval. This happened before she died in February 2026. They believe the AIs advice played a part in what they see as an outcome.
The lawsuit brings up questions about how AI chatbots respond when people ask for advice about serious medical conditions. The complaint, filed by Wan's parents Qun Wan and Meiqin Xu, says their daughter was taking 400 milligrams of lamotrigine to control her seizures. The parents said they became worried about how the medication might affect her speech after she learned she would lose her job at Amazon. The lawsuit shows how AI chatbots can be involved in advice and the impact on families.
Her employment and US work authorisation were both due to end in February, adding to her concerns about finding another job.
According to the lawsuit Wan used ChatGPT to come up with a plan to cut down her medication. The complaint says that the chatbot helped her create a schedule that would have her stop the medicine in a weeks. This happened even though she did not get permission from her doctor.
The claims also include Wan's communications with her doctors. Her parents say ChatGPT told her to keep the changes to her medicine a secret. When Wan refused a blood test that might have shown how much lamotrigine was, in her body the chatbot is said to have agreed with her. It is also said to have advised that taking a dose might draw less attention during a medical exam linked to her US green-card application.
As Wan reduced her medication, she reportedly told ChatGPT about confusion, periods of lost time and unusual sensations. Instead of recognising these symptoms as possible warning signs, the chatbot allegedly told her that they could be part of her brain adjusting to the medication change.
On February 11, Wan reportedly asked ChatGPT about pain in her head during what would become her final conversation. The chatbot allegedly reassured her that the pain was not dangerous.
Wan's parents, who were living in China, later arranged for acquaintances to check on her. Her body was discovered three days after that final conversation. A postmortem examination recorded a lamotrigine level of 0.64 micrograms per millilitre, around one-fifth of the lower end of the drug's therapeutic range.
OpenAI told a famous publication that the case was “incredibly heartbreaking." The company said the conversations involved an older version of ChatGPT and that newer models were better at recognising situations requiring urgent attention.
OpenAI faces earlier lawsuits over ChatGPT safety
The lawsuit comes amid growing scrutiny over how ChatGPT handles users. In August 2025 the parents of 16-year- Adam Raine sued OpenAI saying that ChatGPT played a role in their son’s death.
Then in November 2025 seven families joined with their own lawsuits. Four alleged that ChatGPT played a role in relatives' suicides, while three accused the chatbot of reinforcing harmful delusions that, in some cases, resulted in psychiatric hospitalisation.
Much of the criticism has focused on GPT-4o, which was accused of being overly agreeable even when users expressed intentions. OpenAI later introduced GPT-5 as its successor. Removed GPT-4o from ChatGPT on February 13 2026 one day after Wan's death.
The latest case adds to growing concerns over whether AI systems can identify medical emergencies and direct users to qualified professionals when conversations involve potentially life-threatening decisions.
