Two of the largest teachers unions in the country have teamed up with Microsoft on a new plan to protect students from the risks of artificial intelligence in classrooms. The American Federation of Teachers, the United Federation of Teachers, and Microsoft announced the agreement on September 9, 2026, in New York.
The plan is called the National AI Safety & Privacy Standard. It sets rules for how AI tools can and cannot be used in schools.
AFT President Randi Weingarten, UFT President Michael Mulgrew, and Microsoft Vice Chair and President Brad Smith made the announcement together. They said the agreement comes because no federal or state government has created similar rules yet.
Under the standard, tech companies cannot use student data to train AI systems. Students cannot be tracked using AI tools either.
The agreement also requires human oversight for any AI decision that affects a student. Companies must explain how their tools work in plain language that parents and educators can understand.
School districts across the country can add these protections directly into their Microsoft contracts. That makes the rules legally enforceable rather than just guidelines.
Weingarten said the agreement has "real teeth" because it can be enforced through contracts. She said the goal was to act instead of just getting frustrated with the lack of government rules.
She pointed to three main goals of the agreement. These are protecting student privacy, giving schools control over AI tools, and giving parents clear information about how the tools work.
What the Agreement Covers
The privacy section says student and teacher data cannot be sold, repurposed, or used to train AI models. Schools keep control over how their data is used, stored, and deleted.
The safety section requires AI systems to include strong security features. It also requires the systems to avoid manipulative or harmful experiences for students.
The transparency section means families must be told what data is collected and how it is protected. Schools also keep the right to decide when and how AI tools are used in classrooms.
This new standard builds on other recent moves to limit screens and AI in schools. New York City Public Schools and Mayor Zohran Mamdani adopted a similar framework last week.
The Los Angeles Unified School District and the New York State United Teachers have also adopted similar frameworks already.
Reactions From Education Leaders
Mulgrew, who leads the 200,000 member UFT, said the agreement gives school districts the tools to fight back against companies that misuse student data. He said districts can now end contracts and seek damages if companies break the rules.
Weingarten also referenced a May 2026 speech where she called for a full screen ban from kindergarten through second grade. In that speech, she also called for banning student facing AI in elementary schools and banning companion chat bots for students under 16.
The AFT first released a set of AI guardrails for schools back in 2024. That document outlined nine principles meant to keep students and families informed.
The union also created the National Academy for AI Instruction last year. The academy trains teachers on how to use AI tools safely and spot misuse.
Weingarten said existing privacy laws like HIPAA and FERPA were not built with AI in mind. She said tech companies now need to take responsibility for the products they build for schools.
Microsoft said it plans to offer this same agreement to school districts nationwide going forward.
