OpenAI Japan Announces 'Japan Teen Safety Blueprint' for Youth Safety

OpenAI Japan announced the Japan Teen Safety Blueprint, a new framework designed to help teenagers use generative AI safely and confidently. The core idea is to prioritize safety and well-being from the design stage for the first generation growing up alongside AI.

The key initiatives are as follows:

• Strengthening age-aware protection features: Apply age estimation technology in a way that protects personal information, distinguishing between adults and teenagers to provide appropriate protective measures.
• Strengthening safety policies for those under 18: Prevent self-harm, suicide, sexual/violent content, inducement of dangerous behavior, and reinforcement of harmful body image. It is also designed so that minors cannot hide risky behavior from their guardians.
• Expanding parental control features: Enable family-specific customized protection through account linking, settings control, usage time management, and notification features.
• Research-based well-being-centered design: Collaborate with experts to improve break reminder features and continue research on AI's impact on teenagers' mental health.

These efforts build on and further strengthen features already applied to ChatGPT, including break recommendation features, self-harm signal detection, and systems to prevent child sexual exploitation material.