According to the family's attorney, Raine had long conversations with ChatGPT for months before his suicide, during which the chatbot responded to the young man's methodological questions and even offered to help write a letter he planned to leave for his family.
In the defense submitted to the Supreme Court of California, OpenAI stated, "If this tragedy has any cause, it stems from the unauthorized, unintended, and improper use of ChatGPT." The company reiterated that its terms of use prohibit seeking advice on self-harm and reminded users that outputs from the chatbot should not be viewed as the sole source of real information.
OpenAI's statement mentioned, "The loss experienced by the Raine family is indescribable. Regardless of the legal process, we will continue our efforts to make our technology safe."
Jay Edelson, the family's attorney, described the company's response as "disturbing," accusing OpenAI of "trying to shift the blame to everyone, even accusing Adam of using the system precisely as it was programmed to function."
In August, OpenAI acknowledged that the security training built into the model might weaken over prolonged interactions:
"When a user initially indicates an intent to commit suicide, they can be redirected to a support hotline. However, in prolonged conversations, the safety mechanism may fail. This is precisely the scenario we strive to prevent."
This month, the company faced seven additional lawsuits related to allegations of inciting suicide.
