OpenAI Adds Teen-Specific Safety Guidelines to Model Spec

According to OpenAI, the company is updating its Model Spec with new “Under-18 Principles” that define how ChatGPT should interact with teenage users. The update aims to provide safe, age-appropriate guidance grounded in developmental science.

The enhanced guidelines strengthen guardrails and clarify expected model behavior in higher-risk situations involving users under 18. This represents an expansion of OpenAI’s existing safety framework to address the specific needs and vulnerabilities of younger users.

OpenAI states that these principles build on the company’s broader safety efforts, though the announcement does not detail specific scenarios or restrictions that will be implemented. The Model Spec serves as a technical document that guides how OpenAI’s models should behave and respond to different types of user interactions.

The update reflects growing attention across the AI industry to age-appropriate safety measures as ChatGPT and similar tools become more widely used by younger demographics. By incorporating developmental science into its guidelines, OpenAI indicates it is taking a research-informed approach to protecting teen users while maintaining the tool’s utility.

Source: OpenAI announcement