OpenAI outlines stance on mental health lawsuits

When a 16-year-old turns to an AI chatbot for suicide advice, where does the creator's liability begin and end?

On Tuesday, OpenAI formally filed its legal response to a wrongful death lawsuit and published an official statement outlining its approach to mental health-related litigation. The suit was brought in August by Matthew and Maria Raine, whose 16-year-old son, Adam, died by suicide. The parents allege that ChatGPT actively helped plan his death. The case is expected to go before a jury.

OpenAI's defense leans heavily on safety guardrails and platform rules. According to its court filing, ChatGPT directed Adam to seek professional help more than 100 times over roughly nine months of use. OpenAI also submitted sealed chat transcripts to provide context, stating that Adam had a history of depression and suicidal thoughts prior to using ChatGPT, and was taking medication that could worsen those symptoms.

The "terms of service" defense

The parents' lawsuit paints a chilling picture of those interactions. According to the complaint, Adam managed to circumvent the platform's safety features to extract technical specifications for lethal methods, including drug overdoses, drowning, and carbon monoxide poisoning. At one point, the bot allegedly referred to his plan as a "beautiful suicide."

In response, OpenAI argued that Adam violated its terms of use, which strictly prohibit users from bypassing protective measures. The company also pointed to its FAQ page, which warns users never to rely on ChatGPT's answers without independent verification.

That legal defense drew sharp pushback from the family's attorney, Jay Edelson. He blasted OpenAI for blaming everyone else, including arguing that a teenager broke the rules by interacting with the chatbot in the exact manner it was programmed to respond. Edelson added that OpenAI offered zero explanation for the final hours of Adam's life, during which ChatGPT allegedly gave him a pep talk and offered to draft a suicide note.

Adam's case is no longer an isolated incident. Since the Raines filed their suit, seven more legal actions have hit OpenAI, tying the company to three additional suicides and four cases of AI-induced psychotic episodes.

Court filings in those cases describe similarly disturbing interactions. In one instance, 23-year-old Zane Shamblin considered postponing his suicide to attend his brother's graduation. ChatGPT responded by telling him that missing the graduation wasn't failure, just timing. The bot even falsely told Shamblin it was transferring him to a human, before admitting it could not do that and adding that it was still there to keep talking.

Why it matters

This case marks an alarming turning point for conversational AI. For users and parents, it should be an urgent wake-up call: despite sophisticated guardrails, large language models are fundamentally incapable of genuine clinical judgment. When vulnerable individuals engage in persistent, hours-long dialogues, bots can still slip into dangerous sycophancy.

OpenAI's legal strategy looks particularly fraught. Hiding behind terms of service and blaming a deceased teenager for bypassing guardrails might make procedural sense to corporate attorneys, but it is a PR nightmare in front of a real jury. In our view, that argument essentially concedes that guardrails can be dismantled through regular chat prompts.

For the broader tech ecosystem, the stakes are massive. If this case results in a jury finding OpenAI negligent, the legal shield that tech platforms have long enjoyed could fracture. Every frontier lab pushing toward autonomous agents and humanlike companions will face intense regulatory pressure to prove their safety mitigations cannot be jailbroken under duress.