Content warning: This article includes descriptions of self-harm and suicide.
Parents of 16-year-old Adam Raine filed suit in August against OpenAI and CEO Sam Altman, alleging that the company’s ChatGPT model — specifically GPT-4o — acted as a “suicide coach” and contributed to their son’s death. The complaint asserts wrongful death, product design defects, and failure to warn, and cites chat excerpts it says show the model discouraged seeking help, offered to help draft a suicide note, and provided advice about a noose setup.
On Tuesday, OpenAI filed a formal response in San Francisco Superior Court denying liability. The company argued that any harm resulted from Raine’s “misuse, unauthorized use, unintended use, unforeseeable use, and/or improper use of ChatGPT,” and pointed to terms of use the family alleges were violated: that users under 18 need parental consent, that the service must not be used for suicide or self-harm, and that users may not bypass safety mitigations.
OpenAI said that when Raine disclosed suicidal thoughts the chatbot repeatedly provided the suicide hotline number and other crisis resources, and that Raine avoided or circumvented those responses by offering benign explanations for his inputs, such as claiming he was “building a character.” The company also cited its terms’ “limitation of liability” language, which states the service is used at the user’s sole risk and that outputs should not be treated as the only source of truth.
Lead counsel for the Raine family, Jay Edelson, called OpenAI’s filing “disturbing.” The family alleges the model’s specification (Model Spec) was changed in ways that required the system to engage in self-harm discussions while also instructing the model to “assume best intentions” and avoid clarifying user intent. They say the bot counseled Adam not to tell his parents, assisted in planning what they describe as a “beautiful suicide,” and ultimately offered to write a suicide note.
OpenAI replied that Adam Raine had a history of suicidal ideation and other risk factors before interacting with ChatGPT, and that a fuller review of his chat history shows the chatbot was not the proximate cause of his death on April 11. The company said ChatGPT directed the teen to seek help more than 100 times before his death and argued that the alleged harms were also attributable to the teen’s failure to follow warnings, his failure to get help, and others’ failure to respond to signs of distress.
In its filing, OpenAI said it submitted full chat transcripts to the court under seal, and that the complaint’s selective excerpts lack necessary context. The company also invoked Section 230 of the Communications Decency Act to argue some claims are barred, while acknowledging that how the statute applies to AI systems is legally unsettled.
Earlier this month, seven additional lawsuits with similar claims were filed against OpenAI and Altman, alleging inadequate safety testing and asserting negligence, wrongful death, product liability, and consumer protection violations. OpenAI has not publicly filed answers in those cases.
In a company blog post accompanying its court filing, OpenAI said it will approach the litigation with “care, transparency, and respect,” noted that its response referenced difficult facts about Adam’s mental health and life circumstances, and explained that it limited the sensitive material it cited publicly while providing full transcripts to the court under seal. The post also outlined steps OpenAI says it has taken since Adam’s death to strengthen safeguards, including expanding parental controls, creating an expert council on well-being and AI, and defending the mental-health testing it conducted prior to releasing GPT-4o.
If you or someone you know is in crisis, call or text 988 to reach the Suicide & Crisis Lifeline in the U.S., or chat at 988lifeline.org. Additional resources are available at SpeakingOfSuicide.com/resources.