Microsoft has signed a legally binding agreement with the American Federation of Teachers to adopt sweeping privacy and safety standards for its AI tools in K-12 schools. The pact, announced last week, bars the company from using student or educator data to train AI models, selling that data, or using it for advertising. It also prohibits AI features designed to foster emotional dependency in students. The standards will apply to all U.S. school districts with Microsoft contracts starting November 1, 2026.
The deal arrives as the country's two largest school districts - New York City and Los Angeles - imposed yearlong moratoriums on student AI use while they audit existing ed-tech contracts. The debate over what data AI companies collect and how they use it has intensified as chatbots become embedded in classroom tools students are required to use daily.
What the Microsoft-AFT agreement requires
Under the terms, Microsoft will not use student or educator data to train its AI systems, with a narrow exception for data that could help protect students. The company cannot sell collected data, use it for advertising, or apply it to product development. The agreement also prohibits any AI companions or features "designed to foster emotional attachment or dependency" and bars tools that keep students engaged longer than a learning task requires.
The pact mandates third-party audits for compliance and requires Microsoft to disclose to families, in plain language, how its tools work. Brad Smith, Microsoft's vice chair and president, said the company will extend the agreement to every school district across the country. The framework is legally binding.
Pressure on other tech giants to follow
AFT President Randi Weingarten called the agreement the "first of its kind" and said she contacted Google directly before the announcement, hoping the company would sign on. Google has not responded to a request for comment. The company found itself at the center of controversy in August when it turned on its Gemini chatbot for many K-12 students using Google Classroom, which serves 150 million teachers and students worldwide. Until then, Google had restricted Gemini to students 18 or older.
OpenAI and Anthropic both issued statements praising the agreement and confirmed they are in talks with the AFT to finalize their own pacts. OpenAI called the deal an "important milestone for AI safety and privacy in schools." Anthropic said it was working with the union to shape "a gold standard for safety and privacy."
Josh Golin, executive director of online safety nonprofit Fairplay, cautioned that the agreement sidesteps a larger question. "I worry that a high-profile framework like this will be misunderstood and schools will think, 'If they're doing a good job on data privacy and safety, then that means we should be using the tool,'" he said.
Shifting the negotiating burden away from districts
Elizabeth Laird, director at the Center for Democracy and Technology, said the standards could reshape how schools negotiate with AI providers regardless of whether they are Microsoft customers. Until now, a district's ability to secure privacy protections often depended on its financial clout. "I think this standard makes progress toward not having that type of inequity across school districts," Laird said.
She noted the agreement goes "above and beyond" what is legally required of vendors. For instance, Microsoft committed not to use sensitive data such as gender, race, age, and ZIP code that can re-identify individuals even after names are removed. The framework also highlights, Laird said, that the burden of protecting student privacy should fall on tech companies rather than on parents, students, or schools.
Weingarten emphasized that the union wants parents and educators - not ed-tech companies - to control children's education. "That's what this agreement is," she said. The AFT launched its National Academy for AI Instruction for teachers a year ago, backed by $23 million from Microsoft, OpenAI, and Anthropic.
Why this matters for educators
For teachers and school administrators, the agreement provides a concrete benchmark to use when evaluating any AI vendor's privacy and safety commitments. Even if a district is not a Microsoft customer, the standards - no training on student data, no advertising use, no emotional manipulation features, plain-language disclosures, third-party audits - offer a checklist for contract negotiations. The yearlong audits underway in New York City and Los Angeles signal that districts are now scrutinizing data practices that many ed-tech tools have operated under for years. With AI Learning Path for Teachers programs emerging to help educators navigate these tools, understanding which privacy provisions to demand is becoming as critical as knowing how to use the technology itself.
Your membership also unlocks:
