On Tuesday night, the Sarasota County school board in Florida adopted a seven-page AI policy on a 3–1 vote. That’s the headline. The part I keep thinking about came during public comment, when a high school senior stood up and told the board, “Teachers aren’t told what tasks AI should be used for.” That’s according to WUSF’s reporting on the meeting.

He named the gap most districts are living in. A school board AI policy can set guardrails, list approved tools and assign responsibility. It can’t help a seventh-grade science teacher decide, on a Thursday afternoon, whether AI will help students learn on this assignment or quietly do the learning for them.

That decision doesn’t live in a document. It lives in people. And building it takes longer than a board meeting.

Key takeaways

  • Passing a school board AI policy is a starting line, not a finish line. Students in Sarasota told their board that teachers still don’t know which tasks AI is for.
  • Florida now requires districts to adopt AI policies by July 1 and to train teachers and administrators on AI risks, limitations, safety and compliance. Compliance training is a floor, not a plan.
  • New Stanford research suggests the task matters: using AI to offload academic work can reduce learning, while AI used to understand material shows more promise but mixed results.
  • Short trainings rarely change practice. Turning policy into daily judgment takes sustained, embedded support.

A school board AI policy in Sarasota, and why it isn’t just a Florida story

According to WUSF, the policy was a year and a half in the making and was updated to match a new Florida state rule. The lone no vote, board member Beth Mayberry, said there was “no rush necessarily” and wanted more input from students, families and the community. Vice chair Liz Barker voted yes, saying the policy gives the district “some guardrails.” Superintendent Terry Connor said the district needs a better grasp of what it expects from students and will keep revisiting the policy.

Everyone in that room had a point. And a student telling you teachers don’t know what AI is for is about the most useful feedback a board can get.

The state context makes this bigger than one district. In September, Florida’s State Board of Education approved statewide AI rules, as WUSF and Central Florida Public Media reported. Districts and charter schools have until July 1 to adopt and implement AI policies. Teachers and administrators have to receive training on AI risks, limitations, safety and compliance. Schools have to keep a public list of approved AI instructional tools, let parents opt their child in or out of directly using AI tools, and address how teachers can verify student work.

Wherever you lead, you can see where this is heading: more policy requirements, more tool lists, more mandated training hours. None of that is bad. But if we’re not careful, we’ll end up compliant on paper and confused in the classroom.

The real question your policy can’t answer: which tasks?

That student’s comment lines up almost perfectly with new research. On November 10, Stanford’s SCALE Initiative released The Evidence Base on AI in K-12: A November 2026 Review, which grows its causal evidence base from 20 studies to 52. The finding that stuck with me is simple: using AI to offload academic tasks can reduce student learning. Using it to help students understand new material looks more promising, but the review is careful to say that neither type of use consistently improves learning.

For teachers, the review finds AI can save time on lesson planning and improve the amount and quality of feedback. Chalkbeat’s coverage adds a caution worth hearing: in some studies, students who believed a lesson was AI-generated reported being less motivated. Report author Susanna Loeb put it this way: “On their own without thought, they have some negative effects on learning.”

So whether AI helps or hurts often comes down to the task and the teacher’s judgment about it. A policy can name that judgment. It can’t build it. I wrote about this from the leadership side recently in “AI Replacing Teachers Is the Wrong Debate. Ask Which Work.” The same question applies in classrooms: which work are we handing to the machine, and which work has to stay with the student?

Why a compliance training won’t close the gap

Here’s where I get a little provocative. When a state requires AI training, the easiest way to comply is a single session in August with a sign-in sheet. Check the box. Move on.

We already know how that story ends. The U.S. Department of Education’s 2007 review of the research by Yoon and colleagues screened more than 1,300 studies and found only nine that met What Works Clearinghouse evidence standards. In those nine, teachers who received substantial professional development, an average of 49 hours, boosted student achievement by about 21 percentile points. The three studies with the least professional development, 5 to 14 hours total, showed no statistically significant effect on student achievement.

It’s an older study and not about AI, so I hold it loosely. But the lesson is familiar: short, one-time training rarely changes classrooms. And knowing which tasks AI should and shouldn’t touch in this subject with these kids is a skill that develops through practice, coaching and conversation over time.

I’ve been doing long-term partnerships with schools since 2009, starting with 1:1 laptop rollouts. We had acceptable use policies then too. What changed learning wasn’t the paperwork. It was steady support for teachers to figure out what the laptops were for. AI is the same lesson on a faster clock.

What school leaders can do this month

  1. Ask your students what’s unclear. Hold a short listening session with a dozen students and ask: “When are you unsure whether you’re allowed to use AI?” Their answers will beat any survey.
  2. Translate the policy into tasks. Pick one department and list five common assignments. For each, decide together: no AI, AI for support (explaining, practicing, feedback), or AI as a tool for parts of the work. Sort them by the Stanford question: does this help students understand, or does it do the thinking for them? Share the examples, not just the rules.
  3. Plan training as a sequence, not an event. If you’re required to train staff, map it across the year: an opening session, follow-up practice in PLCs, classroom coaching, and a spring check-in on what changed.
  4. Put a review date on the policy. Tools and rules will keep shifting. Schedule a standing review each semester, with student and family voice built in. Our earlier post on who decides when vendor defaults change has a simple process you can borrow.

For a fuller roadmap, see our K12 AI Implementation Guide.

Why this is partnership work

I’ll be fair: your internal team can and should own this. A good outside partner builds that capacity; it doesn’t replace it. But turning policy into everyday judgment touches everyone at once: the board, the leadership team, coaches, teachers making task-by-task calls, students who need clear expectations, and parents who want to know what’s happening. That’s a lot to carry while running a school.

A long-term partnership spreads that work across a year or more. It might start with a board and cabinet session on what the policy actually commits you to, then move into teacher workshops built around real assignments, coaching cycles, a student session on AI and learning, and a parent night. It can be face-to-face, virtual, or a mix. Every contract is different, because every school’s starting point is different. That continuity matters even more when platforms and state rules keep shifting under you. Tricia Friedman and I talked about this shift toward yearlong, multi-year support on the Shifting Schools podcast episode “AI in Schools Is Changing: What Leaders Need Now.”

I bring 25 years in education and an international lens from working with US districts and international schools. One workshop participant summed up what this work is supposed to produce: “It is no longer intimidating. I have a ‘why’ and ‘how’ for the use of AI in the classroom.” That “why and how” is exactly what Sarasota’s students said is missing. Growth like that tends to build over time, the way the Alaska Staff Development Network’s AI series grew from a fall cohort to a spring cohort to advanced sessions on assessment and learning with bots.

Every partnership is tailored, so reach out and we’ll talk through pricing and what a partnership could look like for your school.

FAQ

What should a school board AI policy include?

At minimum: how students and staff may use AI, student safety and data protection, how tools are approved, academic integrity, and a regular review date. Florida’s rule also requires a public approved-tool list and parent opt-in or opt-out.

Is one AI training session enough to meet a state requirement?

It may satisfy a rule, but PD research suggests one-time sessions rarely change practice. Plan a sequence with follow-up practice and coaching.

How do teachers decide which tasks students should use AI for?

Start with the learning goal. Recent Stanford research suggests AI used to offload academic work can reduce learning, while AI used to understand material is more promising but mixed. Teachers need shared examples by subject, not just rules. Our podcast episode “AI Literacy Is Problem-Solving Literacy”, on verification and judgment, is a good staff conversation starter.

The policy is the easy part

Passing the policy took a vote. Making it real will take a year of conversations, practice and adjustments. That’s not a failure of the policy. That’s just how people learn.

So here’s my question for you: if a student stood up at your next board meeting and asked what AI is for in your classrooms, would your teachers all give the same answer?

Find your starting point. Take one of our free AI readiness reflections (leadership, IT or teaching and learning) to see where your school stands, then reach out at jeff@jeffutecht.com to talk through pricing and packaging.

Sources