Responsible AI Report 2025

This is Microsoft's official AI governance document, setting out the rules Microsoft applies to how it builds and uses artificial intelligence across products like Copilot, Azure AI, and Bing. The most important thing for everyday users is that Microsoft commits to keeping humans in control of high-stakes AI decisions and to being transparent about when AI is involved in interactions affecting you. You should review the privacy settings in your Microsoft account to understand what data is used to train or personalize AI features you interact with.

This document is Microsoft's AI governance framework (published June 2025), establishing internal and external standards for the responsible development, deployment, and oversight of artificial intelligence systems across Microsoft products and services. The framework creates obligations for Microsoft to implement human oversight mechanisms, conduct impact assessments, enforce prohibited use policies, and maintain transparency with users and regulators regarding AI system capabilities and limitations. Notably, the document engages voluntary commitments alongside regulatory compliance language, which may create ambiguity about the enforceability of stated obligations — a deviation from more binding governance instruments adopted by some industry peers. The framework directly engages the EU AI Act, GDPR, and emerging US federal AI policy guidance from the NIST AI Risk Management Framework, implicating both EU and US regulatory enforcement regimes. Material compliance considerations include whether Microsoft's internal governance controls satisfy the EU AI Act's requirements for high-risk AI systems, and whether transparency and accountability commitments meet FTC unfair or deceptive acts or practices standards applicable to AI-powered products in the US market.