Microsoft has agreed to ten safety and privacy principles governing AI use in schools, in partnership with the American Federation of Teachers (AFT) and its New York City affiliate, the United Federation of Teachers. The agreement follows growing unease among parents and educators about AI and screens in classrooms, with New York City and Los Angeles both recently introducing one-year bans on AI tools for many students while they work out appropriate safeguards. It marks an attempt by Microsoft to get ahead of further restrictions by offering districts contractually enforceable commitments rather than voluntary promises.

The principles include a pledge not to train AI models on student or educator data, limits on the data Microsoft collects, plain-language disclosures to families about how tools work, a ban on AI companions, and mandatory human review of "high-risk" decisions. From November, school districts will be able to add these terms to new or existing contracts without renegotiating them entirely. AFT president Randi Weingarten, who has previously called for AI to be kept away from students before middle school, described the terms as "iron-clad" and said legally enforceable provisions were needed given the absence of federal action, even as the union separately runs a $23 million AI training programme for teachers funded by Anthropic, Microsoft and OpenAI.