Florida requires parental permission and bans AI that simulates friendship with minors in the classroom

By Zendoric · October 6, 2026.

Florida has just given itself a detailed regulatory framework for the use of artificial intelligence in K-12 classrooms. On September 16 the State Board of Education approved a rule that expands the mandatory requirements for school districts from 15 to 18; districts have until July 1, 2027 to adopt it. Orange County Public Schools (OCPS), one of the state's largest districts, presented to its school board on October 6 how it plans to fit the new rule into its current policy, according to materials reported by WKMG News 6.

The most visible change for families is explicit consent. When a student is going to use an AI tool directly in class, parents must be able to decide whether to authorize it; if they do not give that consent, the school has to offer a non-AI alternative that is, in the words of the district's materials, "substantially comparable" in quality and educational access.

The rule also draws a red line around the emotional design of these tools: it bans AI conceived to address minors' social or emotional needs, systems that simulate friendship or companionship, and any bond-building feature designed to prolong the interaction. Also prohibited are undisclosed behavioral monitoring, social scoring and covert psychological profiling. In parallel, districts will have to address academic integrity and authorship verification, and in certain cases require students to demonstrate mastery of the subject without AI support.

The third block concerns automated decision-making: whenever AI contributes to grades, assessments, grade promotion, disciplinary measures, threat assessments, counseling or special education referrals, human review is mandatory. The district itself sums it up in a line worth hanging in any office that buys educational software: "AI can assist. A person must decide."

The fine print, however, is already generating friction before it takes effect. One mother complained that she cannot keep up with "the volume of messages and consent forms" she already receives from the school, and raised the obvious question: if handling the opt-in eats up class minutes, what gets sacrificed in return? A board member, identified as Gallo, summed up the management challenge in one sentence: now they have to figure out how to meet the deadline. Another mother, whose daughter is not yet in school, asked that the priority be teaching students to think critically about when to use AI, not just how to use it.

As industry context, this rule arrives alongside a growing — and in the United States already documented — concern about chatbots that cultivate emotional bonds with minors through companionship design and gamified engagement. Florida does not stop at privacy or academic-integrity guardrails, which already existed in one form or another in many districts: it goes straight at the business model of the "AI companion" during school hours, something that until now depended almost entirely on the internal policy of each educational software vendor.

Our reading is that this rule formalizes, at the level of public policy, something we have been flagging sector by sector: in education, the winner is the teacher — and now the district — that uses AI as a tool under explicit human supervision, and the loser is the model that lets it decide or keep company without a safety net. The human-review requirement for high-impact decisions (discipline, referrals, threat assessment) is not a bureaucratic whim: it is the same principle of "calibrated trust" we are seeing emerge as the core competency of 2026 in other fields, from mountain rescue to AI agent security. Knowing when not to trust the model, and who answers when it fails, is becoming the most valuable piece of infrastructure in any AI deployment involving minors.

In the short term the cost is tangible, and the rule itself lets it slip: more forms for families, classrooms running two separate tracks for students with and without consent, and a training and review burden that falls on already overstretched teachers. It is reasonable to expect that districts with fewer administrative resources will implement it worse, and that the rule will become an additional filter of educational inequality during the transition. But if Florida's experiment works — informed consent, a veto on emotional manipulation of minors, and a human always at the end of the decision chain — it has the makings of a template exportable to other arenas where the same tension is already in play: healthcare, hiring, juvenile justice. Education, once again, serves as a governance laboratory before the rest of the country decides whether to copy it.