Microsoft and teachers union agree on enforceable AI privacy safeguards for schools
Microsoft and the American Federation of Teachers have signed a legally binding agreement that prevents the company from using student or teacher data to train AI models. The deal also bans tracking students and requires human oversight for AI decisions in schools. Microsoft will be liable for breach of contract if it violates the terms, with protections taking effect November 1.
The pact, effective November 1st, makes Microsoft contractually liable for any violations, though narrow safety and security exceptions exist for data usage. The AFT, representing the nation's second-largest teacher workforce, framed this as a necessary response to the absence of comprehensive federal AI regulations.
This development occurs amid a wave of local restrictions, including recent one-year bans on AI for younger students in New York City and for all public school pupils in Los Angeles. The union is also pursuing parallel agreements with other major AI developers like OpenAI and Anthropic.
This agreement could establish a precedent for AI deployment in education, potentially shifting power toward institutional oversight. It may prompt other districts to demand similar contractual safeguards, while encouraging vendors to design privacy-first features to avoid liability. The impact could be a more cautious, human-in-the-loop integration of AI in classrooms, though enforcement and scalability remain open questions.