OpenAI adds open source tools to help developers build for teen safety

On Tuesday, OpenAI announced the launch of a new suite of prompts aimed at helping developers ensure the safety of their applications for teenage users. This initiative includes a set of safety policies that can be integrated with OpenAI's open-weight safety model, gpt-oss-safeguard. Instead of starting from scratch to enhance safety measures, developers can utilize these prompts to strengthen their creations. These guidelines tackle critical issues such as graphic violence, sexual content, harmful body image portrayals, risky behaviors, and age-restricted products. By providing these safety policies as prompts, OpenAI ensures that they can be seamlessly adapted for use with other models, although they are likely to yield the best results within OpenAI's ecosystem. OpenAI collaborated with AI safety advocates, including Common Sense Media and everyone.ai, to develop these prompts. Robbie Torney, Head of AI & Digital Assessments at Common Sense Media, stated, "These prompt-based policies help set a meaningful safety floor across the ecosystem, and because they’re released as open source, they can be adapted and improved over time." The AI lab acknowledged that many developers, even those with ample experience, often find it challenging to translate safety objectives into clear, actionable rules. This difficulty can lead to inconsistencies in implementing safety measures or overly broad filtering, which can leave gaps in user protection. OpenAI emphasized the importance of establishing clear and well-defined policies as a foundation for effective safety systems. While the new prompts do not solve all the complexities surrounding AI safety, they build on previous efforts, including parental controls and age prediction features introduced last year. However, OpenAI has faced criticism as well, including lawsuits from families of individuals who died by suicide following extreme use of ChatGPT, highlighting the ongoing challenges in ensuring user safety. Despite these hurdles, the introduction of these safety prompts represents a significant step forward, particularly for independent developers seeking to create safer AI-powered applications.