Microsoft releases Humanist Code of Conduct for AI development

Microsoft published the first draft of its Humanist AI Code of Conduct, a set of values and guardrails for its AI models. The document prioritizes human control and safety, and prohibits AI use in weapons, offensive cyberattacks, deepfakes, and violent content. The final version will be released at the end of the year and will guide model development from 2027.

• Human control and safety prioritized over all other goals
• Prohibited: weapons, offensive cyber, deepfakes, violent content
• Models must obey pause, redirection, and shutdown
• AI declared a tool, not a person; no rights or well-being