Microsoft agrees to stop using school data for AI training
A new contract with a major US teachers union legally bars Microsoft from using student and teacher data to train its AI models, filling a gap left by the lack of federal regulation.
Microsoft has entered into a legally binding agreement with the American Federation of Teachers that explicitly prohibits the company from using data collected from students or educators to train artificial intelligence models. This move marks a significant shift in how school technology is governed, moving away from voluntary commitments to enforceable legal standards.
The primary driver behind this deal is the current absence of comprehensive federal laws regulating AI in education. By establishing specific privacy and safety rules through contract law, the union aims to create immediate protections for families and staff. As reported by Engadget, this agreement serves as a practical workaround to a regulatory vacuum that has left many school districts without clear guidelines on how AI tools operate.
Enforceable privacy standards replace vague promises
Unlike previous industry guidelines that could be easily ignored, this contract includes specific liabilities. If Microsoft violates the terms, it faces breach of contract claims, giving the union and participating school districts real legal leverage. The agreement specifically bans the tracking of students and prevents AI products from making decisions in educational settings without human oversight.
The deal also mandates transparency, requiring Microsoft to clearly explain how its tools work to parents and teachers. This addresses a common concern among educators who feel left in the dark about how their data is processed. The union argues that these provisions are necessary because general corporate privacy policies are often insufficient when dealing with sensitive student information.
Safety exceptions and human oversight requirements
While the ban on data training is broad, the agreement carves out narrow exceptions for safety and security scenarios. This means Microsoft can still use data in specific, limited contexts to ensure the immediate safety of school environments. However, these exceptions are tightly defined to prevent them from becoming a loophole for broader data exploitation.
A key component of the new standards is the requirement for human oversight. AI systems integrated into schools cannot independently make decisions that affect students. This ensures that teachers and administrators remain in control of educational outcomes, mitigating the risk of algorithmic bias or automated errors impacting a child's academic record.
Broader push for school AI regulations
This agreement is part of a wider movement to control AI adoption in schools. Several major US cities, including New York and Los Angeles, have recently implemented moratoriums on the use of AI for younger students. These local actions reflect growing concern among parents and educators about the pace of technological integration in classrooms.
The American Federation of Teachers is now in discussions with other major AI developers, such as OpenAI and Anthropic, seeking similar enforceable protections. If successful, these talks could establish a new baseline for how AI companies interact with the education sector, setting a precedent that prioritizes student privacy and safety over data collection.
