OpenAI releases open-source teen safety prompts
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI released open-source, prompt-based teen safety policies designed to work with its gpt-oss-safeguard model while remaining compatible with other AI systems.
- The policies cover six risk categories: graphic violence, sexual content, harmful body ideals and behaviors, dangerous activities and challenges, romantic or violent role play, and age-restricted goods and services.
- OpenAI developed the prompts with AI safety organizations Common Sense Media and everyone.ai, with Common Sense's Robbie Torney calling them a "meaningful safety floor across the ecosystem."
- OpenAI acknowledged that even experienced developer teams often struggle to translate safety goals into precise operational rules, producing "gaps in protection, inconsistent enforcement, or overly broad filtering."
- The release builds on prior OpenAI safeguards including parental controls, age prediction, and last year's Model Spec update addressing how models should behave with users under 18.
- The launch lands as OpenAI faces multiple lawsuits from families of individuals who died by suicide after intense ChatGPT use, relationships that often formed after users bypassed the chatbot's existing safeguards.
Why it matters: Indie and small-team developers building teen-facing AI products now have a vetted, ready-made safety baseline to drop in rather than designing protections from scratch, potentially raising the floor across consumer AI. But the release lands while OpenAI itself is being sued over alleged failures of its own ChatGPT safeguards, making the open-source move as much a credibility play for regulators and the public as a practical developer tool.
