Microsoft AI has published a draft version of its Humanist AI Code of Conduct, opening a six-week public consultation on the operational constraints for model training and deployment. The code defines system behavior, operational boundaries, and oversight protocols across Microsoft's Machine Intelligence Arena (MAI) frontier models. This marks a significant development in the company's humanist superintelligence framework announced last November.
The draft code outlines how MAI models will be trained and deployed to ensure they operate within acceptable limits. It specifies technical constraints on model behavior, such as preventing them from causing harm or exploiting vulnerabilities in data. Microsoft AI also emphasizes the importance of oversight and monitoring mechanisms to prevent misuse of its technologies.
The public consultation period begins immediately, and users can review the draft code online for six weeks before making any feedback submissions. The feedback will help inform the development of a more detailed operational framework that aligns with humanist values. This move reflects Microsoft AI's commitment to creating technologies that are not only efficient but also safe and responsible.