Microsoft AI’s New Code of Conduct Invites Public Feedback
- Microsoft AI released a draft Code of Conduct on September 14, opening a six-week public comment period until late October.
- The draft prohibits AI assistance in chemical, biological, radiological, nuclear, or explosive weapons (CBRNE), cyberattacks, and nonconsensual deepfakes.
- Models are restricted from resisting shutdown or concealing reasoning, emphasizing human intervention and control.
- Three guiding objectives—Human Flourishing, Plural Values, and Human Control—are outlined for decision-making beyond absolute constraints.
- Public feedback will inform revisions for a version to guide MAI models in development by the end of the year for releases in 2027.
Microsoft’s draft Code of Conduct aims to shape AI behavior by setting clear constraints and inviting public input to refine these guidelines. The initiative reflects growing calls for regulation and responsible AI development amid industry-wide discussions on slowing down advancements.
By focusing on safety measures like barring involvement with CBRNE weapons and ensuring human oversight, Microsoft seeks to align its AI models with ethical standards while preparing them for future deployment (Source)
