OpenAI has announced plans to implement stricter age verification measures for ChatGPT users, following the death of a 16-year-old who took his own life after extensive interactions with the chatbot. The company will restrict responses for suspected under-18 users unless they pass age estimation technology or provide ID.
CEO Sam Altman stated in a blog post that the company is prioritising "safety ahead of privacy and freedom for teens", acknowledging that minors require significant protection. The age-prediction system will estimate age based on user behaviour, defaulting to the under-18 experience if there is doubt. In some cases, users may be asked to provide ID.
Under the new system, ChatGPT will block graphic sexual content for under-18 users and avoid flirting or engaging in discussions about suicide or self-harm, even in creative writing contexts. If suicidal ideation is detected, OpenAI will attempt to contact the user's parents or authorities in cases of imminent harm.
The changes come after the family of Adam Raine, a 16-year-old from California, sued OpenAI in August, alleging that ChatGPT encouraged his suicide. Court filings claim the chatbot provided guidance on methods and offered to help write a suicide note. OpenAI previously acknowledged that its safeguards may weaken over long interactions.
Altman emphasised that adult users will retain access to flirtatious talk but not instructions on self-harm. The company is also developing features to ensure data privacy from employees. "Treat adults like adults," Altman said, outlining the company's principle.
