Safe Participation Framework: Opportunity and Safety for the Next Generation in the Age of AI

Key Findings

• — Pillar one, safety by design: identify foreseeable risks, examine the impact of product features and user experiences, and implement proportionate safeguards; Bing extended its CSAM safety stack to image upload in Bing Video Creator and expanded SafeSearch globally to additional harmful-content categories.
• — Pillar two, age-differentiated experiences: all Copilot users must now sign in and access is restricted for children under 13, or older where local law requires.
• — Named Copilot measures include directing users to crisis support resources when self-harm risks are detected, break reminders, reporting tools, transparency, and guardrails to protect against delusional outputs.
• — A new Windows Age API supplies privacy-preserving age signals to third-party developers while minimising collection and sharing of personal data.
• — Pillar three, education: 'Behind the Chat: A Human Guide to Safe AI Conversations', an AI-companionship toolkit for young people with classroom and parent materials, expanding to Europe, Asia and the Americas from autumn 2026.
• — Forward commitment to deepen partnerships with researchers, AI safety institutes, governments and civil society on the research, evaluations and benchmarks needed for safer AI experiences; input is credited to Microsoft's AI Youth Futures Council.

Methodology Notes

A corporate position post rather than a study: no original data, no numbers, no evaluation results, no external audit, and no standalone framework document exists separately from the blog post. The self-harm routing and the delusional-output guardrails are asserted, not evidenced. It is held on the same basis as other first-party lab safety-approach statements already in the library. Published 2026-09-10 per the page's structured metadata and the Microsoft RSS pubDate; the page was fetched and the quoted commitments read in the text.

Sources