Microsoft Agrees to Strict AI Privacy Limits for US Schools

Microsoft has committed to adopting strict privacy standards for its artificial intelligence tools used in American schools. The agreement, negotiated with the American Federation of Teachers, prohibits the company from using student data to train AI models or sell that information for advertising. This move comes as educators and parents grow increasingly concerned about the role of chatbots in classrooms.

The pact is legally binding and requires third-party audits to ensure compliance. It specifically bans AI features designed to create emotional dependency or keep students engaged longer than necessary for a task. While Microsoft says the rules will apply to all its school contracts starting next month, critics note that the agreement does not address whether AI should be present in schools at all.

Data protection terms defined

Under the new rules, Microsoft agrees not to use data from students or educators to improve its AI systems. There is a narrow exception for data that helps protect student safety. The company also cannot use this data for product development or commercial purposes. Families must receive plain-language disclosures explaining how the tools function and what information is collected.

The agreement targets a specific type of AI interaction that has raised alarm among experts. It prohibits features intended to foster emotional attachment or dependency. This includes chatbots designed to act as companions rather than just educational assistants. The goal is to prevent tools from exploiting the vulnerability of young users for engagement metrics.

Industry response remains mixed

Microsoft Vice Chair Brad Smith stated that the standard sets a high bar for child privacy. He said the company will extend these protections to every school district in the country. However, the impact of this move depends on whether other major tech companies follow suit. OpenAI and Anthropic have said they are discussing similar safety pacts with the teachers union, but Google has not announced comparable protections.

Josh Golin of the online safety nonprofit Fairplay noted that the agreement sidesteps a fundamental question: whether AI products belong in classrooms at all. He warned that schools might misinterpret the privacy improvements as a justification for wider adoption. This is a significant trade-off, as stronger privacy does not necessarily mean the technology is appropriate for young learners.

School districts pause AI use

Two of the largest US school districts, New York City and Los Angeles, have announced a one-year moratorium on student use of AI. They are using this time to study how to best incorporate the technology into education. Both systems plan to conduct extensive audits of their education technology contracts with a focus on data privacy and accountability.

The ban in New York City applies to elementary and middle school students, while Los Angeles extends the restriction through high school. These actions reflect a broader backlash against screen time and AI integration in education. The Microsoft agreement, reported by GN technics/ai (en-US), is seen as a step toward industry standards in the absence of federal legislation.