Microsoft has made its new school AI privacy standard available to every school district in the United States, turning a set of voluntary AI safety principles into terms districts can write into their contracts. The National AI Safety and Privacy Standard, negotiated with the American Federation of Teachers (AFT) and the United Federation of Teachers (UFT), reached its scheduled availability date on November 1, 2026.
The standard was announced on September 9 as a 31-page memorandum of agreement between Microsoft and the National Academy for AI Instruction, an educator-led training body set up by the AFT. Microsoft says districts can add its terms to new or existing Microsoft agreements, and providers must make the protections available within 90 days when a district asks for them.
What the standard bans
At its core is a ban on using student or educator data to train AI models, with only a narrow exception for safety purposes. Third parties cannot use collected data for advertising, product development or sale, and schools keep control over how data is used, how long it is kept and how it is deleted.
The standard also bars AI companions and engagement features designed to foster emotional attachment or keep students using a tool beyond their learning tasks. Student monitoring and tracking features are likewise prohibited, unless a school gives explicit written approval. Every automated decision affecting a student must have human oversight, and schools must receive plain-language disclosures about how the AI works.
Compliance is backed by third-party audits, a 72-hour deadline to report a data breach to the school, and a 180-day deadline to delete data from active systems. The agreement runs for two years, then renews in writing.
Enforceable, not aspirational
What sets the standard apart is that it is contractual rather than a law or a pledge. Districts that find violations gain remedies including ending the contract and seeking damages, a deliberate contrast with voluntary guidelines that carry no penalty.
AFT president Randi Weingarten said when the agreement was announced that anything less than legally enforceable provisions would be simply a wish list. Microsoft president Brad Smith described the standard as setting a high bar for child privacy and AI safety.
The scope has limits. It applies to AI products designed and marketed for education; general-purpose productivity and workplace tools are expressly excluded. And the protections do not apply automatically everywhere, a district must incorporate them into its own Microsoft contract, so adoption will vary from place to place. The agreement also leaves blank signature spaces for other AI providers to join.
Why it matters now
The standard arrives as AI tools spread rapidly through classrooms, raising questions about what happens to children's data once it enters a model. Training bans and deletion deadlines address the fear that a student's schoolwork could quietly become fuel for future systems.
It also lands alongside a wider wave of AI rules: California has separately enacted AI laws covering workplaces and health care, most taking effect in 2027. Those are laws, while Microsoft's school standard is a contract framework, but both point in the same direction, toward enforceable limits rather than promises.
The test now is uptake: how many districts write the terms into their contracts, how audits play out in practice, and whether rival AI providers sign on. For now, the standard gives schools something they did not have before, a contract they can point to when a student's data is at stake.



