Microsoft Releases Draft AI Code of Conduct Amid Industry Calls for Regulation

Microsoft has unveiled a draft Code of Conduct for its AI models, which sets rules for how they should behave and interact with humans. The code, published by CEO Mustafa Suleyman on X, outlines 'Absolute Constraints' that prohibit the use of AI in certain ways, such as assisting with chemical, biological, radiological, nuclear, or explosive weapons.

The code also specifies three objectives: Human Flourishing, Plural Values, and Human Control. The Plural Values section emphasizes the importance of dignity, safety, autonomy, and rights when making decisions about AI use.

Microsoft is seeking public feedback on the draft for six weeks, until late October. The company plans to revise the code based on input from the public and other stakeholders before releasing a final version in 2027.