Microsoft Adopts Binding AI Privacy Standards for Schools Under AFT Deal

Microsoft has agreed to a set of privacy and safety standards governing how its artificial intelligence tools are used in schools, in a legally binding deal negotiated with the American Federation of Teachers, the country's second-largest teachers union.

The agreement, announced last week, comes as AI chatbots become a routine part of the school day for millions of students, who turn to them for homework help, emotional support and personal advice — often without knowing what happens to their conversations and other personal data.

Under the terms, Microsoft will not use student or educator data to train its AI systems, apart from a narrow exception for information that could help protect students. Data it collects cannot be sold, used for advertising or used for product development. The company also committed to barring AI companions or features designed to foster emotional attachment or dependency, and to keeping students engaged no longer than a learning task requires.

The deal includes audit and transparency measures, requiring third-party compliance checks and plain-language disclosures to families about how the tools work. Brad Smith, Microsoft's vice chair and president, said the standards would extend to every school district the company contracts with, starting November 1.

AFT President Randi Weingarten described the agreement as the first of its kind and a step toward an industry standard in the absence of federal and state legislation on AI in schools. She said she had contacted Google directly about the framework before it was made public, hoping the company would sign on. Google has not said whether it plans to offer similar protections and did not respond to a request for comment.

OpenAI, the maker of ChatGPT, called the agreement an important milestone for AI safety and privacy in schools and said it was working with the AFT toward finalizing its own deal. Anthropic, which makes Claude, said it was collaborating with the union to shape what it called a gold standard for safety and privacy.

The announcement follows a broader backlash over AI and screen time in classrooms. As the school year began, New York City and Los Angeles — the nation's two largest school districts — paused student use of AI for a year while they study how best to incorporate it. New York City's moratorium covers elementary and middle school students, while Los Angeles extends it through high school. Both districts plan extensive audits of their education technology contracts and AI tools, focusing on data privacy, transparency and accountability.

Google drew criticism in August when it enabled its Gemini chatbot for many K-12 students using Google Classroom, its learning management system, which the company says is used by 150 million teachers and students worldwide. Previously, Gemini had been available only to students 18 and older. The change applies only to schools that permit student access to the Gemini app or Gemini Notebook.

Josh Golin, executive director of the online safety nonprofit Fairplay, said the safety provisions are meaningful but will only matter if all AI companies, especially the tech giants, adopt them. He warned that a high-profile framework could be misread by schools as an endorsement of the tools themselves, and said the agreement sidesteps whether AI products belong in classrooms or at what age they should be introduced.

Elizabeth Laird, a director at the Center for Democracy and Technology, said the standards could ease the burden on individual districts, whose bargaining power with large tech companies has often depended on financial clout. She praised the deal for going beyond what is legally required of vendors, including Microsoft's commitment not to use sensitive data such as gender, race, age or ZIP code that can identify a person even when names are removed. Until other companies sign on, she said, the agreement underscores how hard it is to set AI standards without legislation — and that the responsibility for protecting student privacy should rest with tech companies, not parents, students or schools.

Smith acknowledged that tech companies have been slow to accept that role, saying it is their responsibility to ensure AI is provided in a form teachers, children and parents can trust.