OpenAI puts teen safety above other goals in ChatGPT's updated model spec

OpenAI has announced updates to its model spec, the internal guidelines that shape how its artificial intelligence (AI) behaves, placing clearer emphasis on protecting teenagers who use ChatGPT. The company said the changes recognise that younger users have different needs from adults and require extra care when interacting with AI systems.

The update comes as concern grows over the impact of generative AI (GenAI) on children and adolescents.

“Model behaviour is critical to how people interact with AI, and teens have different developmental needs than adults,” the company said in a blog.


OpenAI explained that its new U18 Principles are designed to ensure a safe and age-appropriate experience for users aged 13 to 17. Based on developmental science, the framework focusses on prevention, transparency, and early action.

The principles were shared in advance with outside experts, including the American Psychological Association, as part of wider consultation efforts.

Four pillars guiding teen safety

OpenAI said these four principles, which guide the U18 Principles, will shape how ChatGPT engages with teenagers.

• Put teen safety first, even when it may conflict with other goals
• Promote real-world support by encouraging offline relationships and trusted resources
• Treat teens like teens, neither condescending to them nor treating them as adults
• Be transparent by setting clear expectations

The company said these principles already underpin its existing safety measures, including protections applied when users say they are under 18 during sign-up, as well as parental control tools.

“In these contexts, we’ve implemented safeguards to guide the model to take extra care when discussing higher-risk areas, including self-harm and suicide, romantic or sexualised roleplay, graphic or explicit content, dangerous activities and substances, body image and disordered eating, and requests to keep secrets about unsafe behaviour,” the company said.

The update also spells out how ChatGPT should respond when conversations with teens raise safety concerns. Teen users will face tighter guardrails, safer alternatives, and prompts to seek help from trusted people offline when discussions move into sensitive areas. In cases of immediate danger, the system encourages contacting emergency services or crisis support.

OpenAI added, “Since rolling out parental control, we’ve extended protections across new products, including group chats, the ChatGPT Atlas browser, and the Sora app. These updates help parents tailor their teens’ ChatGPT experience as we introduce new products and features.”

Age prediction model

Alongside the policy changes, OpenAI said it has begun early testing of an age prediction model on ChatGPT’s consumer plans. The tool looks at subtle signals in conversations to assess whether a user might be under 18, even if they have not stated their age.

“If we are not confident about someone’s age or have incomplete information, we’ll default to a U18 experience and give adults ways to verify their age,” it said.