NEW YORK, United States – Microsoft has agreed to legally enforceable privacy and safety protections for artificial intelligence used in U.S. schools, establishing contractual restrictions on how student and educator data can be collected and used as school systems face growing pressure over classroom AI.
Microsoft announced the National AI Safety & Privacy Standard on Sept. 9 with the American Federation of Teachers and its New York City affiliate, the United Federation of Teachers. Beginning Nov. 1, school districts will be able to incorporate the protections into new or existing Microsoft agreements without renegotiating their contracts.
The agreement covers privacy, safety, transparency and human oversight. Microsoft says the protections are contractually enforceable if the company fails to meet its commitments.
Student data faces new restrictions
Under the standard, Microsoft will not use student or educator data to train its AI systems, with a limited exception involving information that could help protect students. The agreement also restricts the use of school data for advertising, selling information and product development.
The standard also requires Microsoft to minimize data collection and provide schools, educators and families with clear information about how AI tools operate and what information they collect.
The approach puts contractual obligations around protections that have often been handled through broader privacy policies or voluntary commitments.
AI companions and high-risk decisions face restrictions
The agreement prohibits AI companions or features designed to encourage emotional attachment or dependency among students.
It also requires human oversight for high-risk decisions involving students and educators. Microsoft has committed to auditing and transparency measures intended to allow schools to understand how AI systems are being used and what safeguards apply.
■ Related News
The provisions reflect concerns that AI systems used in schools could move beyond educational assistance into areas involving behavioral monitoring, student profiling or decisions with significant consequences for children.
The agreement does not prohibit schools from using AI altogether. Instead, it establishes conditions that districts can attach to their contracts with Microsoft.
Districts can enforce the commitments through contracts
Microsoft says every U.S. school district will be able to adopt the protections beginning Nov. 1.
Districts can add the standard to new or existing agreements without waiting for a contract renewal or renegotiation. The company describes the commitments as legally enforceable, giving districts contractual remedies if Microsoft does not meet the agreed requirements.
That mechanism distinguishes the arrangement from a voluntary industry pledge.
It also gives individual school systems a way to incorporate specific AI privacy and safety requirements into their commercial relationship with a technology provider.
Schools are under growing pressure over classroom AI
The agreement comes as school districts are reassessing how AI should be used by students.
New York City and Los Angeles, the two largest U.S. school districts, have taken steps to restrict or pause student-facing AI tools while examining their educational, privacy and safety implications.
Teachers’ organizations have also pushed technology companies to establish stronger protections for children and educators.
The American Federation of Teachers has sought similar commitments from other AI companies, according to reporting on the agreement. OpenAI and Anthropic have engaged with the union over AI safety and privacy issues, while Google has not adopted an equivalent agreement.
Microsoft move comes amid wider AI safety debate
The school agreement is part of a broader shift in the technology industry toward formalizing safeguards around increasingly capable AI systems.
Microsoft separately released a draft code of conduct in September focused on keeping its AI systems under human control. That framework addresses issues including correction, shutdown and restrictions on harmful behavior.
Other AI companies and technology leaders have also been debating how much development should be constrained by safety requirements. The debate has intensified as AI systems become capable of performing more complex tasks and interacting with external systems.
The school agreement addresses a narrower problem: how AI systems used in classrooms handle children’s data and make or support decisions affecting students.
The industry response remains unsettled
Microsoft’s agreement establishes a concrete contractual framework, but it does not create a nationwide legal requirement for every AI provider serving schools.
Its broader significance will depend partly on whether other technology companies adopt comparable protections and whether school districts demand similar contractual terms from competing providers.
For schools, the immediate change is more specific. Beginning Nov. 1, districts using Microsoft products will have the option of attaching enforceable privacy and safety protections to their agreements, including limits on data use, restrictions on emotionally dependent AI features and requirements for human oversight.
The arrangement puts pressure on other providers to address the same questions, but whether they adopt comparable contractual protections remains unresolved.
Reporting Credit: Microsoft — National AI Safety & Privacy Standard and contractual commitments; American Federation of Teachers and United Federation of Teachers — agreement and education-sector policy position; U.S. school districts — documented AI policies and restrictions.
