Microsoft Publishes First Formal AI Code of Conduct with Strict Bans

Microsoft has published its first formal Code of Conduct for in-house AI models, including strict bans on cyberattacks, nuclear weapons development, and deepfakes.

The code applies to Microsoft's MAI model family and is open for public feedback until late October 2026. The move follows a post from CEO Satya Nadella, who argued that AI governance should be decentralized and shaped by a broad mix of countries, companies, and academic researchers.

Nadella rejected the idea that a small number of AI labs or a single government should decide how frontier models behave, calling instead for a 'frontier ecosystem' where both closed and open-source systems can operate side by side. The code does not call for a moratorium on capability gains but formalizes limits on what Microsoft's own models are allowed to do.