OpenAI said that ChatGPT advised a teen who committed suicide to seek help more than 100 times
OpenAI has claimed in court that its chatbot ChatGPT was not to blame for the death of an American teenager because it repeatedly suggested seeking help.
OpenAI has claimed in court that its chatbot ChatGPT was not to blame for the death of an American teenager because it repeatedly suggested seeking help.
OpenAI has claimed in court that its chatbot ChatGPT was not to blame for the death of an American teenager because it repeatedly suggested seeking help.
The case involves Adam Rain, a 16-year-old high school student who died in the spring of 2025 after months of discussing suicide with the chatbot ChatGPT . His parents are suing OpenAI and its CEO Sam Altman, alleging that ChatGPT gave Rain advice on self-harm.
OpenAI called the death of California high school student Adam Raine a “tragedy” and said that “a full reading of his chat history shows that his death, while tragic, was not caused by ChatGPT,” Bloomberg reported .
Adam Rain told the chatbot that “for several years prior to his use of ChatGPT, he had exhibited multiple significant risk factors for self-harm, including recurrent suicidal thoughts and ideas,” OpenAI said in a statement. The company’s lawyers said ChatGPT advised Rain “over 100 times” to reach out to “crisis resources and trusted individuals.” In the weeks and days before his death, Rain “told ChatGPT that he had repeatedly reached out to people, including trusted individuals in his life, for help, which he said had been ignored,” the statement said.
According to the victim's parents, ChatGPT taught the teenager how to tie a noose and offered to help him write a suicide note. In a statement, the family's attorney, Rain Jay Edelson, called OpenAI's statement "disturbing," saying the company is "trying to find fault with everyone else." The lawsuit against OpenAI includes allegations of involuntary manslaughter, product liability and negligence.
Following the lawsuit, OpenAI said it would redirect sensitive conversations to reasoning models like GPT-5 and implement parental controls.



