< Back to all clusters
[TECHNOLOGY] · United States · 39 sources

started · updated

Microsoft unveils draft code of conduct for Humanist AI

Microsoft has unveiled a draft 37-page ‘Humanist AI Code of Conduct’ to govern the behavior of its future Microsoft AI (MAI) models. The framework, developed over approximately six months, is designed to ensure that artificial intelligence remains under meaningful human control and serves humanity rather than replacing it.

Key provisions of the draft include mandatory requirements for AI models to be interruptible, correctable, and capable of being shut down by humans. The code prohibits models from independently setting goals, expanding their own permissions, or concealing their reasoning processes. Furthermore, Microsoft explicitly rejects the concepts of AI consciousness, legal personhood, or the idea that models are entitled to rights or welfare.

To mitigate extreme risks, the code establishes absolute constraints, prohibiting AI from assisting in the development of chemical, biological, nuclear, or radiological weapons (CBRNE), as well as offensive cyberattacks, nonconsensual deepfakes, and mass-influence operations.

Microsoft AI CEO Mustafa Suleyman announced a six-week public consultation period to gather feedback from experts and the public. The company intends to incorporate this input into a final version by the end of 2026, which will guide model development from 2027 onward.

Entities

Anthropic · Microsoft · Microsoft AI · Mustafa Suleyman · OpenAI · Satya Nadella

Claims

What the coverage asserts, and how many sources carry each claim.

Sources

2 days ago
2 days ago