Terrill Dicki
Jan 17, 2026 01:38
OpenAI introduces new U18 Principles to its Model Specification, establishing age-appropriate AI safety guidelines for teenage ChatGPT users ages 13-17.
OpenAI has updated its Model Specification—the rulebook governing ChatGPT’s behavior—with new “U18 Principles” specifically designed to protect teenage users. The December 2025 update, developed with input from the American Psychological Association, establishes how the AI assistant should interact with users ages 13 to 17.
The move follows OpenAI’s Teen Safety Blueprint unveiled in November 2025 and comes as major tech companies face mounting pressure over youth safety. Meta expanded its own AI safety tools for teens in October 2025, signaling an industry-wide shift toward age-differentiated AI experiences.
Four Core Commitments
The U18 Principles rest on four pillars: prioritizing teen safety even when it conflicts with other goals, promoting real-world support and offline connections, treating teenagers appropriately rather than as children or adults, and maintaining transparency about the AI’s limitations.
ChatGPT will now apply heightened caution when discussions with teen users venture into high-risk territory. This includes self-harm, romantic or sexualized roleplay, explicit content, dangerous activities, substance use, body image issues, and requests for secrecy about unsafe behavior.
“APA encourages AI developers to provide developmentally appropriate protections for teen users of their products,” said Dr. Arthur C. Evans Jr., the organization’s CEO. He emphasized that human interaction remains crucial for adolescent development and that AI use should be balanced with real-world connections.
Technical Implementation
OpenAI is rolling out an age-prediction model across consumer ChatGPT plans. When the system identifies an account as belonging to a minor, teen protections activate automatically. Accounts with uncertain or incomplete age data will default to the U18 experience until adult status is verified.
Parental controls now extend to newer products including group chats, the ChatGPT Atlas browser, and Sora. The company has also partnered with ThroughLine to provide localized crisis hotlines within ChatGPT and Sora, connecting users to real-world support when needed.
Broader Industry Context
The update reflects growing regulatory scrutiny of AI interactions with minors. OpenAI’s Expert Council on Well-being and AI, established in October 2025, continues advising on healthy AI use across age groups. A global network of physicians now helps evaluate model behavior in sensitive conversations.
For investors watching OpenAI’s trajectory toward a potential public offering, the teen safety infrastructure represents both a compliance investment and a competitive moat. Companies that establish robust youth protection frameworks early may face fewer regulatory obstacles as AI oversight tightens globally.
OpenAI stated it will continue refining these principles based on new research, expert feedback, and real-world application data.
Image source: Shutterstock
Credit: Source link
