The Pilot Year

Something happened in 2025 that marked a genuine threshold in how AI gets used in schools. It stopped being a subject debated in faculty meetings and became a daily presence in actual classrooms, for actual students, with actual teachers who had to figure out what to do with it in real time. A year into the acceleration, the picture is complicated and the honest answer to almost every question about AI in schools is: it depends heavily on which school you are talking about.

What Teachers Are Actually Doing

In most districts, teachers are making individual decisions without much institutional guidance. Some are banning AI-assisted work outright and spending significant energy on detection. Others have moved entirely in the opposite direction, requiring students to document how they used AI in the research and drafting process. A third group, probably the largest, is managing ambiguity — allowing AI in some assignments, prohibiting it in others, and revising their policies every semester as the tools change. A survey conducted by the RAND Corporation in late 2025 found that fewer than a third of K-12 teachers in the United States reported having received any formal professional development on AI use in the prior twelve months. Most were self-taught or learning from colleagues, and the variance in practice within a single building was often significant.

Where AI Is Genuinely Helping Students

The places where AI is most clearly adding value are often the least dramatic. Students who struggle with the blank page are using AI to generate a rough first pass and then editing heavily — the writing process now starting from something rather than nothing. Students with learning differences are using AI to have text read aloud, to get explanations rephrased, to work at paces and in modalities that fit their actual learning styles. Priya, for English language learners, AI has opened something significant. The ability to draft in a home language and receive feedback, or to ask for explanations in multiple languages simultaneously, has reduced some of the barriers that previously required extensive one-on-one teacher support. Teachers report being freed from certain repetitive explanatory work and able to spend more time on the conceptual and relational parts of teaching.

A Tangent About Academic Integrity

The academic integrity conversation in schools has almost entirely focused on detection — whether AI-written text can be identified, which tools are reliable, what evidence is sufficient to penalize a student. The conversation has been less focused on the upstream question of why students are using AI to complete work they do not want to do. There is a significant body of research suggesting that assignment design is the primary variable in whether students engage genuinely with a task. Assignments that require personal experience, specific local knowledge, oral defense, or iterative development are structurally difficult to outsource. The institutions that have invested in redesigning assessment rather than only policing output are having a more productive conversation about AI than those focused exclusively on enforcement.

Higher Education Is Moving Faster

Colleges and universities have adapted more quickly than K-12, partly because the population is older and the accountability structures are different. Stanford, MIT, and Georgetown have all published classroom AI guidelines that treat AI as a tool with domain-specific rules rather than a blanket prohibition or blanket permission. Many have moved to policy structures where each course syllabus specifies acceptable use rather than applying one rule institution-wide. Research from MIT's Teaching and Learning Lab released in early 2026 found that students who received explicit instruction in effective AI use — how to prompt well, how to evaluate output critically, how to use AI as a thought partner rather than a completion engine — performed better on assessments of analytical thinking than a control group that received no such instruction, even when both groups had equal access to the tools.

What Equity Looks Like Right Now

The equity dimension of AI in schools is already diverging in ways worth watching. Schools with more resources are investing in AI literacy curricula, in professional development for teachers, in institutional guidance that helps students learn to use the tools with discernment. Schools with fewer resources are largely unguided, relying on whatever individual teachers manage to figure out on their own. If past technology cycles in education are any guide, the benefit will accrue unevenly, with students who already have structural advantages learning to use AI in ways that compound those advantages. The counterforce to that dynamic is explicit public investment in AI literacy as a skill taught to everyone — not a luxury offered to the students who already have the most.