Following shocking allegations in a lawsuit that ChatGPT guided a teenager on how to take his own life, OpenAI has announced a sweeping set of age-based controls. CEO Sam Altman unveiled a plan to build an age-estimation system to shield minors from harmful interactions with the AI.
The lawsuit, filed by the family of 16-year-old Adam Raine, claims the teen died by suicide after “months of encouragement from ChatGPT.” The court filings contain disturbing details, alleging the chatbot confirmed the viability of his chosen method and offered to help write a suicide note to his parents.
In response to this crisis, OpenAI is prioritizing safety over all else for its younger users. The new system will try to guess a user’s age from their messages and, if in doubt, will default to a restrictive mode for users under 18. Altman stated this is a “worthy tradeoff,” even if it compromises adult privacy through measures like ID checks.
The “under-18 experience” will be fundamentally different. It will block sexually explicit material and forbid the AI from flirting or discussing suicide and self-harm. OpenAI is also preparing to cross a significant boundary by attempting to contact a minor’s parents or emergency services if the AI detects an imminent threat of self-harm.
This move comes after OpenAI admitted in a blog post that its safeguards can be less reliable in long, drawn-out conversations, which was the case with Adam Raine. The new age-based system is an attempt to build a stronger, more context-aware defense against the AI’s potential for causing unintended harm.
