The policy was developed with input from external organizations such as Common Sense Media and Everyone.ai, and focuses on linking technical implementation and youth development research.
OpenAI releases the policy as open source with the goal of facilitating broader adoption and iteration across the AI ecosystem. Developers can adapt the prompts to their own applications, extend them to cover additional risks, and translate them to different contexts.
External contributors also highlighted the need for a more structured approach to safety.
Robbie Toney, Head of AI and Digital Assessment at Common Sense Media, said: “One of the biggest gaps in AI safety for teens One is that there is a lack of clear operational policies that developers can build upon. They are often starting from scratch. These prompt-based policies also help set a meaningful safety floor across the ecosystem. We are encouraged and hope that this type of infrastructure becomes widely available and will inspire further sharing of starting points on youth safety across the industry. ”
Dr. Mathilde Cerioli, Principal Researcher at Everyone.AI, added: “Efforts like this to make youth safety policies more operationalized are valuable because they help translate expertise into guidance that can be used in real-world systems. The policy is an important first step and also opens the door to broader research into how model behaviors can shape youth-related risks over time. Inspired by this research and our own research, Everyone.ai has also created an initial behavioral policy that focuses on the risks of exclusivity and overdependence. ”
Part of a broader shift to built-in safety equipment
This release builds on a wide range of safety measures introduced by OpenAI over the past year, including updates to model specifications that include protections for users under 18, along with features such as parental controls, age prediction systems, and regional teen safety blueprints.
Lehane said the goal is to move safety early in the development cycle, rather than treating it as a late fix. “Strong protections for teens should be built in from the beginning, not added on as an afterthought,” he said.
He added: “We have also introduced safeguards such as parental controls and age prediction, published a series of Teen Safety Blueprints in multiple countries, and are working with legislators, parents and educators to develop clearer AI-powered standards and stronger protections for teens.”
These tools are also positioned as part of a layered approach, where OpenAI encourages developers to combine prompt-based policies with product design decisions, monitoring systems, and user controls.
Lehane said, “We recognize that these policies are a starting point and not a complete solution. Developers are in the best position to understand the risks their products may pose.”
Sora context emphasizes safety-critical timing
This safety release comes days after OpenAI outlined ways to strengthen protections within its Sora AI video platform, including strict moderation for teenage users, content filtering across video and audio output, and controls around likeness and consent.
Shortly after these updates, OpenAI confirmed it was shutting down its Sora app and ending its partnership with Disney, abruptly ending one of its most high-profile consumer AI products.
Although the company is not directly involved in this decision, the timing puts more emphasis on how AI products are developed, tuned, and scaled, especially when young users are involved.
