Microsoft has established a set of ten safety and privacy principles for AI use in schools following an agreement with the American Federation of Teachers and its affiliate, the United Federation of Teachers. These terms are designed to be contractually enforceable by school districts that choose to adopt them.
The agreement prohibits Microsoft from training AI models on data belonging to students or educators. It also limits the volume of data collected by the company and mandates that families receive clear explanations regarding the functionality of AI tools. Furthermore, the terms ban AI companions and require human oversight for high-risk decisions.
This move follows one-year bans on student-facing AI tools implemented by school systems in New York City and Los Angeles. Starting in November, school districts will be able to incorporate these new privacy terms into their existing or future contracts without the need for full renegotiation.
The American Federation of Teachers previously announced a $23 million AI training hub for educators, supported by funding from Microsoft, Anthropic, and OpenAI. AFT President Randi Weingarten emphasized that these legally enforceable provisions are necessary to address privacy concerns in the absence of federal regulation.
