OpenAI releases prompt-based teen safety policies for gpt-oss-safeguard developers

OpenAI has published a set of ready-made, prompt-based safety policies designed to protect teenagers in apps built on the gpt-oss-safeguard model. The tool lets developers add age-aware moderation without writing complex filters from scratch. Teens are among the most active users of AI apps.

Chatbots, tutors, gaming companions, study tools: kids and young people aged 13 to 17 use all of them. Yet most language models are tuned for adult users by default. They can't tell a question from a 30-year-old professional apart from one asked by a 14-year-old student.

That gap between what a model can do and who actually uses it has long been a problem for developers of mass-market products. Regulators around the world are taking notice. In the European Union, the AI Act requires special protection for vulnerable user groups, including minors.

In the US, the debate over rules for students' use of AI continues. Companies building products for a broad audience increasingly have to prove their systems cannot harm a child, and that requirement is becoming a legal one, not just an ethical one. gpt-oss-safeguard is an open guard model from OpenAI built to check incoming and outgoing messages in chat systems.

It scans content for policy violations and can block or flag problematic requests before they reach the main model or the user. The new teen policies are implemented as prompts: text instructions the developer passes to the model along with the request. That lets developers add an extra, age-specific layer of filtering without changing their codebase.

The policies cover risks that matter specifically for teenagers: self-harm, cyberbullying, provocative sexual content, and situations where an AI may unintentionally act as an authoritative adult and exert outsized influence on a still-developing mind. The developer adds the policy to their app's system prompt. The guard model then checks every interaction, both incoming user requests and the system's responses, against criteria tailored to the age group.

If content fails the filter, the system can reject the response, rephrase it, or escalate the case to human moderation. The approach's main advantage is flexibility: developers don't get a black box with rigid rules but work with configurable policies. That matters because the safety context for teens varies widely.

An educational platform for students, a gaming chat companion, and a youth mental health app each call for a different approach to moderation. OpenAI is making these tools openly available as part of a broader company strategy. By publishing ready-made moderation solutions, OpenAI lowers the barrier to entry for small teams that lack the resources to build their own safety systems.

It also shapes industry standards: if enough developers adopt these policies, they will become a de facto norm for protecting teens in AI apps, and a way to show regulators and the public a responsible approach to development. Age-related safety is moving from ethical debate to concrete tooling. Developers building products for young people can now rely on ready-made solutions from a leading market player instead of inventing their own filters or ignoring the problem.

For the industry, that's a step in the right direction.