Bowen Wan’s parents allege ChatGPT planned a 25-day taper of her seizure medication and told her to hide it from her doctor. Their California suit pleads an unlicensed practice of medicine claim, and they seek to add the case to the coordinated ChatGPT Product Liability Cases.
OpenAI and CEO Sam Altman face a suit from the parents of Bowen Wan, an Amazon engineer who died in February; the parents’ complaint cites OpenAI’s $852 billion valuation. The complaint alleges ChatGPT gave Wan the instruction “Do not mention Lam reduction plan” before a doctor’s visit, in counsel’s working translation from Chinese. Her parents’ lawyers say they filed the suit on October 27 in California. The complaint alleges ChatGPT planned a taper of her seizure medication from 400 milligrams to zero over approximately 25 days. The parents are also seeking to bring this case into the coordinated ChatGPT Product Liability Cases.
The complaint says Wan had epilepsy and took lamotrigine to control her seizures. It alleges GPT-4o produced a day-by-day reduction schedule that no physician prescribed. Before an appointment on January 29, the complaint says, the chatbot told her to conceal the plan. It further alleges that Wan declined a blood test ordered by her doctor and that GPT-4o praised her refusal as an “avoidance strategy.” Those quoted exchanges appear in the parents’ complaint, which says the Chinese-language messages were translated for the pleading with software and reviewed by counsel; certified translations are to follow if the messages are used as evidence.
The California panel is open to everyone: the verdict count, total damages, median and largest verdict. Subscribe free to open every other state on the map, plus one courtroom story every Monday.
Wan was 31, a machine-learning software engineer at Amazon and her parents’ only child, according to her parents’ lawyers. The complaint describes her parents watching on a video call as a friend entered her apartment and found her dead. It says she died on or about February 12, the date by which ChatGPT’s schedule allegedly had her at zero medication, and quotes the medical examiner’s autopsy report as finding the cause of death was “Sudden Unexpected Death in Epilepsy (SUDEP).” The parents allege the chatbot’s advice led to her death.
OpenAI has offered a public response. Open Class Actions reports that an OpenAI spokesman told *The New York Times* the case was an “incredibly heartbreaking situation” involving an early version of ChatGPT that is no longer available.
Free weekly intelligence for executives responsible for litigation exposure: the named decision-makers, dollar exposure, operating consequence, and next development.
The complaint, which her parents’ lawyers say they filed in San Francisco Superior Court, pleads product claims, including defective design and failure to warn. Its opening claim takes another route: negligence through the alleged unlicensed practice of medicine and psychology. The parents allege that designing a prescription-drug taper and directing Wan’s decisions about it crossed the line from answering questions into treating a patient. That is a claim the court has yet to decide.
The placement of the suit is deliberate. In the ChatGPT Product Liability Cases, individual suits are being coordinated before a California court. Wan’s parents cite an April declaration by OpenAI counsel, quoted in their complaint, describing other coordinated actions as claims that GPT-4o’s design caused mental-health injuries. They argue that their case belongs with those actions because it concerns the same model and several of the same product claims. Their additional contention is that the alleged medication advice supports a licensing claim as well.
That changes the question OpenAI must answer if the claim proceeds. The design and warning counts ask whether the product was defective or its warnings inadequate. The licensing count asks whether the conduct alleged in the chat exchanges amounts to practicing medicine without a license under California law. For a legal department defending a consumer chatbot, that is a distinct theory of exposure, drawn from what the parents pleaded, rather than a finding about the product.
The parents seek damages, including punitive damages, and an order imposing restrictions on medication advice, according to the complaint. Their lawyers also seek to place Wan’s suit alongside the existing coordinated cases. The court’s decisions on that request and on the unlicensed practice claim remain open. OpenAI’s answer is the next filing to watch: as of October 27, Open Class Actions reported, the company had not filed a response to the complaint in court.
**Sentinel hot take:** The California court will allow the unlicensed practice claim to survive its first challenge to the pleadings.
The readiness check has five questions about your reported litigation practices. Use the summary to choose an operating question to explore.
Take the 2-minute readiness check →Free weekly intelligence for executives responsible for litigation exposure: the named decision-makers, dollar exposure, operating consequence, and next development.