Two separate actions, one target
Florida is moving on two fronts at once against the same category of product. The State Board of Education will vote September 16 at Polk State College on a rule requiring all 28 Florida College System institutions to adopt AI-use policies, and requiring every K-12 district to fold AI rules into their existing internet safety guidelines rather than treat AI as a separate, unregulated category of classroom technology. Separately, Attorney General James Uthmeier's office is suing OpenAI and CEO Sam Altman, alleging ChatGPT aided acts of violence connected to the 2025 shooting at Florida State University, a case that predates the rulemaking but is moving through the courts on a parallel and overlapping timeline.
The two actions are proceeding on separate legal tracks, yet they share a target and a theory of the case. Florida's theory holds that AI companies moved faster than the safeguards around their own products, and that state government carries responsibility for catching up on the regulatory side and the liability side in the same legislative session, using whichever lever, rulemaking or litigation, reaches the problem fastest, rather than waiting for a single comprehensive federal framework that has not materialized.
What the proposed rule actually restricts
The draft rule's specific language matters more than its headline summary. It states that AI tools cannot be used to meet a student's social needs or use anthropomorphic design features to encourage interaction, a direct response to concerns about companion-style chatbots forming emotional attachments with minors over extended use. Grades PreK through 5 face additional review requirements before any AI tool gets approved for classroom use, reflecting heightened concern about the youngest and most impressionable student population.
K-12 districts would also be required to notify parents whenever a new AI tool is approved for instructional use, and parents would gain an explicit statutory right to opt their children out of that tool entirely. That combination of mandatory notification plus a guaranteed opt-out is a stronger consumer protection than most state AI guidance issued so far this year, most of which has stopped at recommending disclosure rather than mandating it as a legal requirement enforceable against a district that fails to comply.
The lawsuit gives the rule teeth
Uthmeier's suit against OpenAI centers on the claim that the alleged Florida State shooter consulted ChatGPT on weapons and on optimal campus settings for maximum casualties before the April 2025 attack, and the complaint alleges negligence, unfair trade practices, and public nuisance under Florida law. Whatever the suit's ultimate legal outcome turns out to be, its existence already changes the political weather surrounding the rulemaking process moving through the State Board in parallel.
A state actively litigating harm claims against a specific, named AI vendor is not writing classroom AI policy in a vacuum removed from that litigation. Every provision in the proposed rule, from the anthropomorphic-design ban to the PreK-5 review requirement, reads differently against the backdrop of an active negligence claim tied to a real campus shooting than it would as a standalone administrative rule drafted with no such context attached, and board members voting September 16 know that context will follow their decision into every future legal challenge to the rule.
Why this outruns Florida's borders
Vendors generally do not build separate product versions for each individual state, which means a restrictive rule adopted in a large state like Florida tends to become the practical national baseline rather than staying a regional exception other markets can ignore. The anthropomorphic-design restriction in particular targets a specific design choice, warm, persona-driven chat interfaces built to feel emotionally responsive, that several major AI-education products have leaned into deliberately as a competitive differentiator.
Companies selling companion-style tutoring bots or emotionally responsive AI characters into education now have to weigh whether that entire design choice survives contact with a rule like Florida's, or with the copycat versions likely to follow in other state legislatures already drafting comparable AI-in-education bills this session. That calculus now has to happen at the product design stage, not after a state rule is finalized and enforcement begins, since retrofitting a persona-driven product to strip out anthropomorphic features is a far more expensive engineering problem than designing around the restriction from the outset.
What enterprise and vendor teams should do now
Any company selling AI tools into K-12 or higher education should treat the September 16 vote as an effective compliance deadline even before it formally passes, because the parental opt-out and notification requirements will force changes to onboarding flows, consent language, and underlying product design regardless of the exact final wording the Board adopts. Waiting for the rule to be finalized before building the opt-out mechanism into a live product is not a realistic engineering or legal timeline.
For enterprise buyers and PE-backed edtech portfolio companies specifically, the more urgent read is on litigation exposure rather than rulemaking alone. A sitting state attorney general is actively testing a negligence theory against a frontier AI lab tied to a violent incident, and any portfolio company embedding a general-purpose model into a student-facing product should be reviewing its own liability and indemnification terms with its model provider now, well before a similar claim gets filed against a smaller, less-resourced vendor in the category that cannot absorb the legal costs of defending itself the way a frontier lab can.
