Microsoft's School AI Deal Gives the AFT Too Much Say

Microsoft has agreed to a set of enforceable “AI safety and privacy” standards developed with the American Federation of Teachers (AFT) and its affiliate. School districts can add the standards to contracts for covered Microsoft education AI products. This raises questions about why one of the country’s most politically active teachers unions should have a role in setting standards that could shape students’ ability to use AI. Its president, Randi Weingarten, spoke at the 2024 Democratic National Convention, underscoring how closely the union’s leadership is involved in partisan politics.

The agreement calls these rules “AI Safety” but that phrase can obscure what is actually being decided. Rather than being measures taken to prevent catastrophic events, the deal uses “AI safety” as a buzzword for conforming to the union’s leftist views and promoting “equity” in education. 

Its “equity and accessibility” provision reads like a DEI policy for school AI. It singles out race by requiring providers to test whether their tools perform “materially worse for students of color”. The provision treats race as a possible marker of differences in how an AI tool performs for students. It assumes children of certain races are less capable of using AI, a demeaning premise that has no place in the classroom. Students should be treated as individuals, not presumed to need different standards based on race. If a tool has a specific flaw, Microsoft should fix it. A difference in results between racial groups is a correlation, not proof that race caused the difference, and it should not become grounds for withholding a useful tool from every student like the agreement mandates it does. 

The Trump administration has made AI education a priority. Its April 2025 executive order calls for students to develop AI skills and for teachers to receive training to use AI in the classroom. The Microsoft agreement should be judged in light of that goal: districts need to protect student privacy while giving students meaningful opportunities to learn and use AI. 

School districts should reject this DEI agreement. Privacy protections for students are necessary, but they do not require adopting race-based criteria for evaluating AI tools or giving the AFT-backed Academy a role in withholding approval. Districts should demand strong privacy terms directly from Microsoft while keeping decisions about classroom AI accountable to parents and local school boards.Microsoft has given the AFT a prominent role in setting standards for school AI, despite the union’s partisan record and an agreement that leaves their power to suspend covered tools open to serious criticism.