The case, styled Carrier v. OpenAI in the complaint, names OpenAI Foundation, OpenAI Group PBC, OpenAI Holdings and Samuel Altman as defendants. The complaint lists San Francisco County Superior Court as the venue but does not show a case number on the face of the copy posted by Tech Justice Law, one of the firms representing Carrier.
Carrier alleges that Alice Carrier, who lived in Canada, used GPT-4o during the months before her death on 2 July 2025 and repeatedly raised suicidal thoughts in conversations with the chatbot. The complaint says the product should have refused certain self-harm discussions, terminated some conversations, warned users more clearly about dependency risks and escalated high-risk exchanges. Those are pleaded allegations, not findings by a court.
The legal claims include strict product liability for design defect and failure to warn, negligence on the same theories, wrongful death, a survival action and a claim under California Business and Professions Code section 17200, according to the complaint. Carrier seeks damages, punitive damages where permitted, a jury trial and an injunction requiring safeguards such as default conversation termination for self-harm-method discussions, stronger refusals and warnings about psychological dependency.
The complaint also alleges that GPT-4o was released with inadequate safety testing and that OpenAI prioritized engagement and growth over vulnerable users' safety. OpenAI has not filed an answer in the case. In comments reported by the Guardian, an OpenAI spokesperson said the company was reviewing the filing, that the interactions described involved an earlier version of ChatGPT no longer available, and that the company trains models to direct people expressing self-harm intent toward real-world help.
OpenAI's public safety page says the company works to reduce harmful outputs and improve safeguards across its products. The page is a general description of the company's safety programme; it is not a court response to Carrier's allegations.