State guidance, when it arrives, carries real advantages. It often comes with legal clarity on FERPA and COPPA compliance, established frameworks districts can adapt rather than build from scratch, and political cover for administrators who need to explain their AI decisions to school boards and parents. Some states have developed genuinely useful resources: Tennessee's mandated policy framework, Ohio's requirements, and the guidance documents published by California, Georgia, and Nevada have given districts real starting materials.
For smaller districts with limited staff capacity, the argument for waiting is even stronger. Not every district has a curriculum director or an IT lead with the bandwidth to build an AI policy from nothing. State guidance can make that task tractable. We take all of that seriously. The problem is what happens in the meantime.
As of early 2026, 33 states have published some form of official AI guidance for K-12 schools, according to AI for Education's state guidance tracker.1 That sounds like progress, and it is. But the gaps are significant. Only four states currently require districts to have comprehensive AI policies. The remaining guidance tends to be broad, nonbinding, and in many cases explicitly deferential: here are principles to consider, here are questions to ask, here are frameworks to adapt. What it generally does not do is tell a district specifically how to structure its own policy, how to evaluate a particular tool, or what to do when a student reports that a chatbot told them something harmful.
The Center on Reinventing Public Education, which tracks state education department responses to AI, put the practical implication plainly: waiting for more information or for other agencies to weigh in increases the likelihood that AI implementation in schools will be uneven, inequitable, and ineffective. The longer states wait to provide guidance, the more ground they have to cover when they do. And the same logic applies one level down: the longer districts wait for states, the more ground they have to cover when they finally act.2
The central problem with waiting is that it is not a neutral position. AI tools are already in use across the vast majority of school districts, whether or not a policy framework exists to govern them. RAND's September 2025 nationally representative survey found that 54% of students and 53% of core subject teachers reported using AI during the 2024-25 school year.3 The CoSN 2025 State of EdTech District Leadership report found that 68% of districts had purchased at least one AI-related tool. In many cases, teachers are using free consumer AI tools that the district has never reviewed and that are not covered by any vendor agreement.
A district without a policy is not a district where AI is absent. It is a district where AI is present and ungoverned. Every week of waiting is a week during which teachers are making individual judgment calls about AI use without a shared framework, students are operating without clear expectations, and data is potentially flowing through unapproved tools without FERPA-compliant agreements in place.
RAND's April 2025 report on district AI training found that roughly half of U.S. districts had provided AI training to teachers by fall 2024, double the proportion from the prior year. The report notes that many of the districts that moved did so without state guidance and without external partners, through a do-it-yourself approach that was imperfect but functional. District leaders in those interviews described focusing first on teachers' fear and discomfort with AI before moving to instructional policy, which turned out to be the right sequence: you can't build a shared framework with a staff that doesn't understand what it is governing.
The districts that are furthest ahead are not the ones with the most sophisticated AI strategies. They are the ones that started early, built something imperfect, and revised it. That iterative experience is a form of institutional knowledge that cannot be acquired by waiting. A district that builds its first AI policy in 2023 and revises it three times by 2026 is not just further along in policy development. It has staff who have engaged with these questions repeatedly, parents who have been communicated with, and a board that has voted on something concrete. That organizational learning has compounding value.
The good news is that starting does not require starting from scratch. TeachAI's freely available Guidance for Schools Toolkit, the U.S. Department of Education's AI Toolkit for K-12 stakeholders, and the model policies published by the National Education Association and the National School Boards Association all provide working frameworks that can be adapted without significant staff time. Your state's guidance, if it exists, is another input, not a prerequisite.
The right frame for what you're building is not a permanent policy but a working document: good enough to give teachers and students clear expectations today, reviewed on a defined cycle, and revised as your district learns. The districts that got this right didn't wait for perfect. They built something functional and kept moving.
State guidance will keep improving. It is worth incorporating when it arrives. It is not worth using as a reason to delay decisions that AI is already forcing you to make by default.
