Microsoft Agrees to AI Privacy Rules for Schools

Microsoft signed onto ten AI safety and privacy principles for K-12 schools this week, striking a deal with the American Federation of Teachers and its New York City affiliate, the United Federation of Teachers. The agreement lets school districts write the terms directly into their contracts starting in November, without renegotiating everything else. Among the commitments: Microsoft won't train its AI models on student or educator data, will limit how much data it collects in the first place, and will explain its tools to families in plain language rather than legal boilerplate. The company also agreed to ban AI companions outright and to require a human in the loop for any "high-risk" decision involving students.

The timing isn't incidental. New York City and Los Angeles both rolled out one-year bans on student-facing AI tools just last week, buying time to figure out what guardrails actually make sense. Parents and teachers in both districts have pushed back hard on AI and screen use generally, and Microsoft's move reads as an attempt to get ahead of more bans before they land. The AFT, which represents roughly 1.8 million members nationally, framed the deal as filling a gap left by regulators.