Microsoft Responsible AI Standard

This document sets out Microsoft's rules for how its AI systems—including Copilot—must be built and used responsibly, requiring equal treatment of all users, human oversight of AI, and protection of personal data. If you use Copilot at work, your organization's existing security rules automatically apply, limiting who can see what Copilot generates. You can manage your privacy preferences at any time through your Copilot Privacy settings.

The Microsoft Responsible AI Standard establishes a defined framework of six core values—Fairness, Reliability and Safety, Privacy and Security, Transparency, Accountability, and Inclusiveness—that govern how Microsoft designs, builds, and operates AI systems across their full lifecycle. The document imposes affirmative obligations including non-discrimination requirements, human oversight of AI outputs, responsible model validation, and deployment of specific AI governance tooling such as the Responsible AI Dashboard. For Copilot specifically, the standard requires that personal data be retained securely, that content access be gated by permission controls, and that all existing organizational security and compliance requirements be inherited when Copilot is used in a workplace context. Legal and regulatory compliance and avoidance of harmful bias are established as explicit requirements for responsible AI use, not aspirational goals.

This document means that Microsoft's AI systems, including Copilot, are built under obligations to treat users equally, avoid discrimination based on personal characteristics, and keep personal data secure. Content generated by Copilot is accessible only to users with the right permissions, so access is not universal. Users have an ongoing, unconditional right to manage their privacy preferences—they can do so at any time through the Copilot Privacy settings.