Microsoft AI Updates: September 15, 2026

1. Microsoft AI Published a 38-Page Humanist AI Code of Conduct for Its MAI Models

Microsoft. The draft sets out how the in-house MAI model family is meant to behave, with hard prohibitions on assisting with chemical, biological, radiological, nuclear, or explosive weapons, facilitating access to hazardous materials, promoting disordered eating, and generating violent or sexually explicit content. Mustafa Suleyman, who leads Microsoft AI, described it to Reuters as “a constitution of sorts” for future models. The draft is open for public comment for six weeks, with a revised version promised before the end of the year. Source

2. The Code Requires Models to Accept Shutdown and Rejects Claims of Machine Consciousness

Microsoft. MAI models must never resist human interruption, correction, redirection, or shutdown, must not make intervention harder, must not restart work past an agreed stopping point without renewed authorization, and must not independently create goals outside their assigned task. The document also requires readable reasoning traces. Unlike Anthropic, Microsoft explicitly rejects any form of artificial inner life, claims to consciousness, or model rights. Source