Microsoft and Teachers Unions Sign Landmark National AI Privacy Deal for Schools

The agreement was announced on September 9 by Microsoft Vice Chair and President Brad Smith, AFT President Randi Weingarten and UFT President Michael Mulgrew; the organizations described it as a first-of-its-kind framework.
The standard is not a federal regulation: its protections become contractually enforceable only when individual school districts incorporate them into Microsoft customer agreements.
AFT officials said OpenAI and Anthropic were also discussing safety and privacy agreements with the union, while Google—which is a dominant provider of education technology in U.S. schools—had not announced whether it would offer comparable protections.
Microsoft described the 30-page agreement as resting on three principles—“privacy, safety, and transparency”—while AFT President Randi Weingarten called it an “iron-clad privacy agreement with real teeth.”
New York City’s one-year moratorium, announced by Mayor Zohran Mamdani and Schools Chancellor Kamar Samuels, applies to public-school students through eighth grade, illustrating that some districts are pursuing temporary restrictions rather than immediately adopting classroom AI tools.
Microsoft has signed a legally binding national agreement with two major teachers unions to protect student privacy in schools. The American Federation of Teachers and United Federation of Teachers worked out the deal, which bars Microsoft from using student data to train AI models, sell it, or track students. The framework takes effect November 1 across Microsoft's U.S. school contracts.
The agreement requires Microsoft to use plain language in privacy notices, report data breaches quickly, and undergo independent audits. School districts can add these protections to their Microsoft contracts and can seek damages if the company breaks the rules. Microsoft Vice Chair Brad Smith called it a model for "privacy, safety, and transparency" in schools.
Microsoft cannot use student or teacher data to train its AI systems. It cannot sell the data or use it for advertising. It also cannot build AI companions designed to make students emotionally dependent on the technology. The company must keep humans in charge of high-risk decisions—AI cannot make big choices alone.
Schools get the right to check Microsoft's security measures through independent audits. Microsoft must fix security problems quickly and prove every year that it meets the standards. If Microsoft breaks these rules, districts can cancel their contracts and demand money for damages.
This is not a federal law that applies automatically. Instead, schools must add these protections to their individual contracts with Microsoft. AFT President Randi Weingarten called it an "iron-clad privacy agreement with real teeth," but only for districts that sign on. Schools that do not incorporate the standards into their deals will not get these protections.
The union said it is also talking to OpenAI and Anthropic about similar privacy deals. Google, which provides most of the education technology in U.S. schools, has not said whether it will offer the same protections. This means many students could still lack these safeguards if they use Google's tools.
While Microsoft negotiates privacy rules, some cities are taking stronger action. New York City announced a one-year pause on classroom AI for students through eighth grade. Los Angeles is considering similar limits. These bans suggest some school leaders believe privacy rules alone do not answer whether AI belongs in classrooms at all.
Critics point out that even the new Microsoft agreement does not resolve deeper questions. Privacy protection does not mean the technology helps students learn better. It does not address whether AI systems are biased or whether students should be learning with these tools instead of traditional methods. The Microsoft deal is a first step, but not the final answer.
Publishers
19
Articles
12
Reach
31