Microsoft Signs a Legally Binding AI Privacy Deal With America's Second-Largest Teachers Union

Key points

• Microsoft committed to ten AI safety and privacy principles with the American Federation of Teachers (AFT) and its New York City affiliate, enforceable by school districts that choose to adopt them.
• The agreement prohibits Microsoft from training AI models on student or educator data and bans AI companions in school settings.
• New York City and Los Angeles both imposed one-year bans on many student-facing AI tools, the backdrop against which this deal was struck.
• School districts can add these terms to new or existing Microsoft contracts from November 2025 without renegotiating the whole agreement.
• The AFT, which also runs a $23 million AI training hub for teachers funded partly by Microsoft, called the provisions "iron-clad" and said the federal government had failed to act.

Two of America's biggest school systems just told students to stay away from AI tools. Now Microsoft is trying to show it can be trusted anyway.

The company struck a deal this week with the American Federation of Teachers (AFT), the second-largest teachers union in the United States, and the AFT's New York City branch, the United Federation of Teachers (UFT). Together they agreed on ten privacy and safety principles that school districts can make legally binding the moment they sign up, as first reported by The Verge.

What exactly did Microsoft promise?

The ten commitments cover four things parents are most likely to ask about. Microsoft pledged not to use student or teacher data to train its AI models. It agreed to collect less data in the first place. It will explain to families, in plain language, how its tools work. And it banned AI companions, software designed to act as a student's friend or confidant, from its school products.

The agreement also requires human review before any "high-risk" decision gets made using AI output. That means a machine cannot, on its own, determine something with serious consequences for a child's school life.

From November, any school district can attach these terms to a new or existing Microsoft contract without having to reopen and renegotiate the entire deal. That is a practical detail worth noting: districts will not need lawyers for months to get these protections in place.

Why is this happening now?

New York City and Los Angeles both imposed one-year bans on AI tools for many students in the past week, giving themselves breathing room to think through proper guardrails. Microsoft almost certainly read those decisions as a warning sign that other districts could follow.

AFT President Randi Weingarten has herself called for a ban on screens before third grade and on student-facing AI before middle school. Yet the AFT also runs a $23 million AI training hub for educators, funded by Anthropic, Microsoft, and OpenAI. The union's position is complicated: sceptical of unguarded AI in classrooms, but unwilling to leave teachers unprepared for it.

In a statement, Weingarten said the terms were "iron-clad" and necessary "because no one else, including the federal government, has stepped up to do the real work." She added that anything short of legally enforceable rules is "simply a wish list."

What does this mean for parents and teachers?

If your school district opts in, these commitments become contractual obligations Microsoft must keep, not voluntary guidelines it can quietly drop. That is a meaningful distinction.

Nothing forces a district to sign up, and the deal covers Microsoft's tools only. It sets no rules for the many other AI products schools might choose. But for the districts that do adopt it, families gain a clearer, enforceable picture of what happens to their children's data inside Microsoft's classroom software.