Florida's Opt-In Rule for School AI Tools Could Cut Off Weakest Readers

Florida's Department of Education wants parents to have the final say over whether their children use AI tools in class. A proposed amendment to the state's Internet Safety Policy, discussed at a rule development workshop on 5 August, would require every district and charter school to let parents opt their children in or out of AI-based instructional tools, and to cap how long an opted-in child spends using them. Districts have until January 2027 to comply.

On paper this looks like straightforward parental control, the kind of consent mechanism UK schools already navigate for photography or off-site trips. In practice, industry groups and some educators warn it could do the opposite of what it intends: strip AI support from the students who rely on it most.

A broad definition, a narrow fix

The trouble starts with the rule's scope. The Software and Information Industry Association (SIIA) told the Department that its draft definition of "artificial intelligence" is so wide it would catch purpose-built, district-vetted instructional software alongside general-purpose chatbots such as ChatGPT or Character.AI. A reading-support programme that adapts text difficulty in real time, or a maths platform that sequences practice questions based on error patterns, would trigger the same opt-in and time-limit requirements as an open-ended companion chatbot.

That matters because those adaptive tools are exactly the ones aimed at struggling readers, English language learners and students with disabilities — the groups Florida's own legislature has spent years trying to better serve. SIIA argues that bundling a blanket opt-in with an alternative-assignment requirement creates a heavy administrative burden for teachers, who would need to track individual permissions and prepare parallel non-AI activities, while disproportionately locking out the pupils for whom the adaptive tool is not an add-on but scaffolding.

The rule was written to address chatbot risk. As drafted, it also governs the assistive software students with additional needs depend on.

Why this happened through rulemaking, not legislation

Florida's legislature failed to pass a bill this year restricting AI companies' interactions with minors, despite it being a stated priority of Governor Ron DeSantis. Rather than wait for another legislative session, the Department of Education is using its existing authority over school internet safety policies — a category that already exists in state law — to impose AI-specific requirements through rulemaking instead. That route is faster but gives affected parties less room to negotiate scope before a rule takes effect, which is part of why SIIA's objection focused on definitional breadth rather than the principle of parental choice itself.

The UK contrast

England's approach under KCSIE 2026 has gone the other way: safeguarding duties sit with the school, not individual parental sign-off per tool, and the emphasis is on risk assessment and staff training rather than consent forms for specific software. Florida's model puts the decision with families on a tool-by-tool basis, which sounds protective until you consider that most parents have no way to judge whether a given piece of software is an open-ended chatbot or a closed, curriculum-aligned support tool. The Department's own workshop drew pushback precisely because the distinction the rule needs to make — general-purpose AI versus vetted instructional AI — is not one its current text makes.

What to do

UK school leaders watching this from a distance should note the underlying lesson, not just the US specifics: any consent or opt-in policy for classroom AI needs a working definition of what counts as "AI" before it goes anywhere near governors or parents. A policy that cannot distinguish a spell-checker with adaptive hints from a general chatbot will end up either meaningless or, as in Florida's case, capable of removing support from pupils with SEN or EAL needs. If your school or trust is drafting AI guidance this term, get your SENCO and EAL lead into that conversation early, not after the policy is signed off.

What to watch

Florida's final rule is due before January 2027. Whether the Department narrows its AI definition in response to SIIA's comments, or leaves the broad wording in place, will be a useful test case for any UK multi-academy trust considering a similar tool-by-tool parental consent model rather than a whole-school risk framework.

Sources:

• Florida schools directed to add AI to internet safety policies
• SIIA Urges Florida to Narrow Proposed AI Rule for K-12 Classrooms
• Florida Weighs Parental Opt-In Rules for AI in Schools
• Florida Department of Education to hold online workshop on AI classroom guidelines
• As students head back to school, districts grapple with new AI policies