On September 14, Microsoft AI rolled out the MAI Model Code of Conduct, laying the groundwork for the development of 'Humanistic AI'. This code was crafted following extensive discussions with experts from diverse fields such as AI, law, ethics, and more. It is now accessible for public feedback over a six-week period. The code underscores the necessity for AI to consistently remain under human supervision, forbids the creation of autonomous consciousness or the replication of human emotions, and imposes strict safety limits. Microsoft intends to publish an updated version by the year's end, which will serve as a guide for model development from 2027 onwards.
