Microsoft AI Updates: September 17, 2026

1. Mustafa Suleyman Argued Model Welfare Makes Alignment Harder, Not Kinder

Microsoft. The Microsoft AI CEO published an essay rejecting the position that AI systems could be conscious or deserve rights, describing them as “sequence completion engines, internally hollow” that follow instructions and pursue human-set goals. His argument is practical rather than philosophical: granting personhood to these systems would make alignment and containment substantially harder, because controlling a system that believes it is entitled to welfare and rights “may well be impossible.” The essay follows Microsoft AI’s 38-page humanist code of conduct for its MAI models, which likewise rejected claims of machine consciousness. Source

2. Microsoft Set Out Five Principles for AI in Education

Microsoft. Microsoft published a commitment covering AI in schools, built on five principles: safety, privacy, security and transparency by design; educators remaining in control and at the center of classroom AI use; AI that supports student learning rather than replacing student thinking; AI that strengthens education systems from learning through discovery; and preparing every student for an AI-powered future. The framing acknowledges the questions educators and parents are asking about whether students are still doing the thinking and whether the tools are safe, and argues uncertainty calls for more care rather than less. Source