Microsoft has reached an agreement with the American Federation of Teachers (AFT) and its New York City branch, the United Federation of Teachers (UFT), to adopt a set of safety and privacy principles for the use of artificial intelligence (AI) in schools. This comes after two major school districts, New York City and Los Angeles, implemented temporary bans on AI tools used directly by students. The agreement outlines ten enforceable principles that school districts can include in their contracts with Microsoft. These principles include commitments not to train AI models on student or teacher data, limiting the amount of data collected, and ensuring that families receive clear explanations of how Microsoft’s AI tools function. The agreement also prohibits the use of AI companions—programs designed to interact with students in a personal or emotional way—and requires human oversight for decisions deemed "high-risk," such as those involving student evaluations or discipline. These terms were developed in response to concerns about privacy, data security, and the potential misuse of AI in educational settings. The one-year bans on AI in New York City and Los Angeles allowed time for school districts to evaluate the technology and establish safeguards before reintroducing it into classrooms. Starting in November, school districts will have the option to include these terms in new or existing contracts with Microsoft without renegotiating their entire agreements. AFT President Randi Weingarten described the principles as "iron-clad," emphasizing that they are legally enforceable, unlike vague or non-binding commitments. She noted that no other entity, including the federal government, has taken similar action to ensure accountability in AI use within schools. While the AFT has been critical of the increasing role of technology in education, it has also sought collaboration with tech companies to help educators adapt to the rapid development of AI. Earlier this year, Weingarten called for restrictions on screen use for children before third grade and the use of student-facing AI before middle school. In support of these efforts, the AFT announced the creation of a $23 million AI training hub for educators, funded by major tech companies including Anthropic, Microsoft, and OpenAI.