On September 9, 2026, the American Federation of Teachers (AFT) announced an agreement with Microsoft in a Manhattan press room attended by approximately 50 reporters and union members. The agreement prohibits Microsoft from tracking student use of its AI tools and using student and teacher data to train AI models. The guidelines apply to any school district that wishes to adopt similar principles when forming their own agreements with Microsoft and purchasing its AI products.
The AFT has given Microsoft and any other tech companies that sign the agreement a 180-day deadline to delete records of AI data use from students and school districts upon request. While Microsoft has agreed not to use student and teacher data for AI training, exceptions exist for “narrow safety and security” situations where minimal data may be used to address security threats or user harm.
The contract includes provisions for school districts to pause or discontinue Microsoft's AI systems if there are material breaches of the agreement. However, the guidelines do not specify clear rules for what constitutes a reasonable timeframe for remediation, leaving it to Microsoft and the school districts to negotiate in good faith.
When asked about accountability measures if Microsoft fails to comply, Microsoft Vice Chair and President Brad Smith stated that the company would publish reports for transparency and adhere to deadlines for fixes. He emphasized the importance of resolving issues without resorting to legal action, stating that the goal is to fix problems as they arise.
Concerns about the feasibility of enforcing the contract have been raised, especially given Microsoft's history of promoting AI products to schools. The announcement follows recent actions by New York City Mayor Zohran Mamdani, who declared a one-year moratorium on AI use for elementary and middle school students, and the Los Angeles Unified School District's restrictions on generative AI platforms for the 2026-27 school year.
The AFT's contract with Microsoft concludes with the statement, "When in doubt, protect the student," but the practical implementation of these AI use policies remains uncertain.