Study: The National AI Policy Landscape in K–12 Education

In the span of just two years, artificial intelligence has moved from an emerging curiosity to an operational reality in American K–12 classrooms. Students are using it to draft essays and build interactive study apps. Teachers are using it to generate lesson plans, differentiate instruction and develop assignments. Administrators are using it to summarize data, create chatbots and organize teams. And most school districts are scrambling to update or create policies that reflect the ever-changing world of AI. But what does an AI policy landscape look like? How many districts have formal AI policies at all? What do those policies say? And what does the distribution of approaches reveal about the state of readiness, equity and strategic thinking in K–12 education? To answer those questions, we built the AI School Policy Database by using a five-level AI policy continuum to code 122 districts and schools across 38 states and then analyzed the results. What we found is a snapshot of a field that is neither panicking nor confidently leading: it is waiting, watching and managing uncertainty one teacher-directed decision at a time.

What This Snapshot Tells Us

The dataset does not capture every district/school in America; it is a structured sample. But the patterns it reveals are consistent and interpretable. Taken together, they point in the same direction.

• Most districts/schools are reactive. The dominant posture is conditional permission, teacher-delegated decision-making and a wait-and-see attitude toward systemic AI integration.

• Some districts/schools are still trying to stop the clock. Nearly 30% are actively restricting or prohibiting AI use, an approach that is increasingly difficult to sustain as student access to AI tools extends far beyond school networks.

• Only a small fraction is daring to lead. Fewer than one-third of districts/schools have published policies that reflect genuine strategic governance: formal frameworks, defined approved uses, staff development and equity-centered design.

• Geography matters. Regional variation in policy progressiveness is substantial and reflects structural differences in state-level investment, board culture and guidance infrastructure.

• State guidance is necessary but not sufficient. Districts/schools in states with official guidance are not reliably more progressive than those without. Legally mandated guidance appears to drive adoption more reliably than advisory frameworks.

• Policies are written for students, not for districts/schools. The dominant framing of AI policy as student conduct governance leaves critical institutional questions about staff use, procurement, equity and organizational strategy largely unanswered.

As we interviewed districts/schools over the past year, it has become clear that more attention must be paid to adults and systems. One of our findings is that policies emphasize students with a focus on cheating — when, in actuality, most of the districts we are working with this year have had greater risks presented by adult staff members misuse, including sharing a PDF copy of the copyrighted curriculum with a personal LLM account and violating copyright agreements; sharing IEP and evaluation information without deidentifying student data. These missteps hold real long-term risk for districts. The ability of our education systems to integrate AI responsibly is critical to our ability to uphold data privacy for staff, students and families alike.

The Five Levels of the AI Policy Continuum

Level 1 – Pro-Innovation/AI Encouraged: Actively promotes AI as a learning tool; embeds AI literacy into curriculum; proactively deploys AI tools and provides open access to students and staff; emphasizes equity and opportunity in AI use

Level 2 – Guided Integration: Has published a formal AI framework or multilayered guidance document that defines approved uses, builds AI literacy and addresses privacy and ethics; has structured AI onboarding for students and staff

Level 3 – Conditional/Teacher Directed: Permits AI use but only when explicitly authorized by the teacher; focuses on citation, academic integrity and limits on substitution; does not have a districtwide AI curriculum; delegates decisions on AI use to individual teachers

Level 4 – Restrictive/Integrity Focused: Tightly restricts AI use or only allows it when teacher gives permission; only allows use of vetted/approved AI tools; places strong emphasis on dishonesty enforcement and penalties when evaluating AI use

Level 5 – Prohibited/No Use: Explicitly prohibits AI use or treats it as plagiarism/academic dishonesty; does not provide a pathway for authorized student AI use; uses discipline or gives students zeroes as consequences for AI use violations

The 122 organizations spanned public school districts, charter schools and individual schools across all four major U.S. census regions. We coded district/school policies based on publicly available documents including board of education policies, acceptable use policies, staff and student handbooks and official district/school websites. We tracked the status of state-level guidance using the Center on Inclusive Design and Digital Learning (CIDDL) as a reference.

Overall, 3.3% of districts/schools operated at Level 1, 27.9% operated at Level 2, 44.3% operated at Level 3, 17.2% operated at Level 4 and 7.4% operated at Level 5.

Most Districts Are in the Cautious Middle

The most striking finding from our dataset is the concentration of districts/schools that operate at Level 3. The idea that AI is permitted, but only when the teacher says so sums up the policy posture of nearly half (44.3%) of the districts/schools in the dataset.

Level 3 policies share a common logic: they do not ban AI outright, but they do not take a strategic position on it either. Instead, they delegate the decision to individual teachers. Students may use AI only when explicitly authorized and are required to attribute AI-generated content as such. Students are prohibited from submitting AI-generated content as their own. The district/school sets boundaries; teachers set the rules within those boundaries.

This approach is understandable given the pace of AI development. But it carries real risks. When AI governance is delegated entirely to individual teachers, equity of access becomes uneven. Students in classrooms with AI-forward teachers get fundamentally different educational experiences than peers in the same building whose teachers remain restrictive. District-level strategy is replaced by classroom-level improvisation. A Level 3 policy often means the superintendents and administrators have started AI governance work but have not yet developed a true strategic framework.

Nearly 1 in 4 Districts/Schools Restrict or Ban AI

Although Level 3 is the most common policy posture, a substantial share of districts/schools (nearly 25%) has moved in the opposite direction: 17.2% operate at Level 4 and another 7.4% operate at Level 5.

Level 4 districts/schools are primarily concerned with academic integrity, emphasizing detecting and prohibiting the submission of AI-generated content. Level 5 districts/schools take it a step further and issue categorical bans on AI tools entirely.

Restrictive policies are not randomly distributed. They cluster in specific states and regions. Among states with no official guidance, districts/schools tend to have more restrictive or variable policy postures. Florida (with an average level of 4.5 across all districts/schools in the dataset), Texas (average of 5.0), Indiana (average of 3.6) and Wisconsin (average of 3.6) have disproportionately high concentrations of restrictive-to-prohibitive approaches.

It is worth noting that restriction and prohibition are not inherently wrong. Some districts/schools have made deliberate, values-based decisions to limit AI for pedagogical reasons. But the findings from our dataset suggest that in many cases, prohibition is a proxy for policy uncertainty — a way of buying time rather than an intentional governance choice.

Active AI Promotion Is Rare

Only 3.3% of districts/schools in the dataset operate at Level 1. These are districts/schools that embed AI literacy into the curriculum, actively deploy tools across the entire district/school, provide students and staff with open and supported access and emphasize equity of access as a policy design principle.

Another 27.9% operate at Level 2, meaning they have published a formal AI framework with guidance on approved uses, staff onboarding and privacy and ethics.

Regional Variation Points to Structural Inequities

The dataset reveals meaningful regional variation in how districts/schools approach AI governance. The Midwest lags notably behind other regions, with an average policy level of 3.2, about half a point more restrictive than the Northeast (2.7). This pattern holds even after accounting for state AI guidance: Midwestern states like Indiana, Wisconsin and Ohio have substantial proportions of Level 4 and 5 districts/schools. The Northeast and West show the most embracing, driven in part by early-adopter districts/schools in California, New Jersey, Oregon, Virginia and Washington. The South is split: some states are producing some of the most progressive district-level policy language in the country, while others are among the most restrictive policies we reviewed.

These regional patterns matter because they are not random. They reflect structural differences in state-level policy infrastructure, edtech investment capacity, school board cultures and the degree to which districts/schools feel supported by state-level guidance. Thus, a student’s relationship with AI in school is, in part, a function of where they live.

State Guidance Matters, But Its Influence Is Uneven

Of the 122 districts/schools in the dataset, 63% are in states with officially published AI guidance from the state department of education. However, only 15.6% explicitly reference state guidance in their own policy documents. This decoupling of state guidance from local policy development is one of the most important findings from the dataset.

States have done significant work: as of early 2025, more than 28 states had published AI guidance for K–12 schools. But that guidance is not reliably making its way into district/school policy. By and large, districts/schools are writing their own frameworks without anchoring to the resources states have developed, or they are not writing frameworks at all.

When state guidance becomes a legal mandate, district- and school-level policy follows. When it remains advisory, districts/school adoption is inconsistent. States that have published guidance but have not mandated adoption should consider what additional supports, such as model policies, technical assistance and tiered implementation tools, would help districts/schools move from awareness to action.

Most Policies Speak to Students, Not Systems

Nearly two-thirds (65%) percent of policy documents in the dataset list students as the primary audience. Only 22% have policies explicitly addressed to both students and staff. Just one organization has a policy focused primarily on staff.

This distribution reveals a gap: most AI policy in K–12 education is framed as student conduct governance, not as institutional strategy. The policies answer the question of what students can do with AI far more often than they address how the district/school should use AI to better serve students.

Policies that focus exclusively on student behavior miss the larger organizational dimensions: How are teachers being trained to use AI responsibly? How is the district/school vetting AI tools for data privacy and algorithmic bias? How is the district/school protecting staff and student data? How is leadership thinking about the role of AI in special education, multilingual learner support, or mental health services? These questions require policy that speaks to the whole organization, not just to student handbooks.

Recommendations for District/School Leaders

Based on our findings, we recommend the following to superintendents and district/school administrators:

• Treat AI policy as strategy, not compliance. A policy that tells students when they can and cannot use generative AI is a starting point, not a destination. Effective AI governance requires district- or school-level strategic thinking: What AI tools will we use? What use cases for AI will we prioritize and systematize? How will we vet tools and their outputs? How will we train our staff? How will we ensure equitable access? These questions belong in board-level policy, not just student handbooks. They necessitate forward thinking commitments that allow for safe innovation by adults.

• Engage your state’s guidance resources. Only 15.6% of districts/schools in this dataset explicitly reference state-level guidance — a missed opportunity. State education agencies have invested significant resources in AI frameworks, model policies and procurement guidance. Districts/schools that are not anchoring to these resources are doing more work than necessary.

• Move prohibition to a transition plan. Districts/schools currently at Level 4 or 5 should honestly ask themselves if prohibition working. Students with smartphones have access to AI around the clock. A policy that prohibits AI on school networks without addressing the broader context of students’ AI access outside of school is not a governance solution. It’s a governance deferral. The question is not whether students will use AI but whether schools will help them learn to use it well.

• Design for access from the start. The Level 1 districts/schools in this dataset recognize that every policy decision either expands or limits students’ opportunities to benefit from AI. Rather than treating equity as a separate initiative, they build considerations such as device availability, internet access, accessibility, multilingual supports, professional learning and approved AI tools into the policy from the outset. As districts/schools develop or revise their AI policies, they should ask explicitly whether each policy decision will create or reduce disparities in student and/or staff access to meaningful AI use. The answer should continually shape the policy.

• Build policy with your community. The most durable AI policies in this dataset emerged from multistakeholder processes: committees of teachers, students, parents and administrators working together. Policy built with the community carries legitimacy that top-down mandates cannot. It also surfaces practical concerns about academic integrity, data privacy and student well-being that benefit from diverse perspectives.

What We Learned

The AI policy landscape in K–12 education reflects a field in genuine transition. The governance conversation has clearly begun. But the distribution of that progress suggests a field that has not yet found its footing.

What the data make clear is that a single policy posture is not the path forward, nor is a single tool or use case. What districts/schools need is a sustained governance process — one that moves from reactive management of student behavior to proactive design of AI-enabled systems, connects local practice to state-level resources and treats access and opportunity as a non-negotiable design principle.

The districts/schools at the leading edge of this dataset offer a proof of concept. They are not waiting for perfect clarity about where AI is headed. They are building frameworks resilient enough to evolve, transparent enough to sustain community trust, thoughtful enough to provide safety and security and flexible enough to ensure that the use of AI by adults can support outcomes for all students.

That may be the model worth scaling.