Analysis: After a public hearing on September 16, 2026, the Florida State Board of Education adopted new rules governing the use of artificial intelligence, including amendments to Rule 6A-1.0957, the Internet Safety Policy rule. The amended rule requires every district school board and charter school governing board to adopt AI-specific amendments to their internet safety policies by July 1, 2027. While the rule’s stated purpose, protecting students from unsafe uses of artificial intelligence, is laudable, several provisions present serious practical and legal challenges. This article examines the most significant concerns and offers actionable guidance for school board employees.

Districts and charter school boards must adopt clear policies governing AI and other educational technologies. The new rules require parents to be notified whenever an AI instructional tool is approved. The notice must identify the tool, the courses in which it will be used, and how students will interact with it. Parents must be given the opportunity to opt their child into using the tool. For students who do not opt in, parents must be offered and select a non-AI alternative.

Tools used in VPK through grade 5 must undergo additional review to determine whether they are appropriate for students’ ages and developmental levels. Policies must prohibit AI designed to address a student’s social or emotional needs, as well as AI that simulates companionship or an emotional relationship. AI also may not use human-like features intended to encourage students to interact with a machine. Undisclosed behavior monitoring, social scoring, and psychological profiling are also prohibited. Vendors may not sell student data or use student data to train commercial AI models. Finally, the policies must establish clear expectations for academic integrity and must ensure that any AI instructional tool deploying autonomous/agentic AI is under the direct supervision of an adult, maintains a complete activity log, and includes a mechanism allowing school personnel to immediately suspend or disable its autonomous functionality without vendor assistance. The policies must also ensure that AI is never used as a decision maker absent human review, including for grading, evaluations, promotions, discipline, threat assessments, counseling referrals, Exceptional Student Education (ESE) referrals, or communications, and require districts to disclose to the Department any unexpected autonomous behavior or IT safety or security incident caused by an approved AI tool, including unauthorized data access.

The Overbroad Definition of AI The rule defines an “Artificial Intelligence system” as “a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments.” This definition is so expansive that it captures virtually every piece of modern software in a school environment. Spell-checkers that predict the next word, search engines that recommend results, adaptive learning platforms, calculator apps with predictive features, and even spreadsheet auto-complete functions all arguably qualify. While the rule’s operative requirements focus primarily on the narrower category of “artificial intelligence instructional tools,” which now includes software applications or services using generative or autonomous/agentic artificial intelligence, including machine learning, made available to students for educational purposes, the broader “AI system” definition still appears in the purpose clause, the definition of “instructional use,” and the prohibitions on social-emotional AI.

The September 18 amendment also defines “autonomous artificial intelligence” or “agentic artificial intelligence” as a machine-based system that operates through a continuous cognitive loop allowing it to be proactive, set goals, make multi-step decisions, execute actions, and gather, log, or generate data sets with minimal or no human oversight. By expressly including these systems within the definition of an AI instructional tool, the amendment further broadens the rule’s reach. This creates a compliance gray area where districts must classify each software tool individually, a substantial administrative burden for districts managing hundreds of software licenses. Without interpretive guidance from the Department of Education, districts risk either over-complying or under-complying and facing enforcement actions.

Parental Opt-In Impracticability Subsection (3)(e)1.e requires that parents be given the option to opt their child into using an AI instructional tool, with a non-AI alternative instructional tool provided for students who are not opted in. This mechanism is operationally unworkable. AI-adjacent technologies are deeply embedded in the platforms Florida schools rely on daily, including Google Classroom, Microsoft 365 Education, and adaptive reading programs such as Lexia and iReady, all of which incorporate machine learning features. If a parent declines to opt in, the district must provide a comparable non-AI alternative, a requirement that may be impossible where AI is integral to the product’s functionality. Districts would face maintaining parallel instructional ecosystems: one AI-enabled and one AI-free. The staffing, licensing, and curricular alignment costs are enormous. The September 18 amendments compound this operational burden for districts using agentic AI tools, which must provide direct adult supervision, maintain complete activity logs, and implement mechanisms to immediately suspend or disable autonomous functionality.

–Hidden Costs of Implementation The agency’s rulemaking notice claims the rule “can be implemented with existing staff and technology” and will not exceed $200,000 in regulatory costs. This is difficult to reconcile with the rule’s actual requirements: –
–Technology Audits. Districts must catalog and classify every software application to determine whether it qualifies as an AI tool, a project spanning instructional platforms, administrative tools, and third-party applications that could take months.
–Staff Training. Subsection (3)(e)9 mandates teacher and administrator training on AI risks, limitations, and compliance, requiring curriculum development and professional development hours.
–Consent Management. The opt-in requirement and detailed parent notification provisions will require systems to track individual student opt-in status across multiple tools, grades, and academic years.
–Recordkeeping and Reporting. Subsections (3)(e)11–13 require maintaining publicly accessible approved-tool lists, 30-day student interaction records, and detailed reporting to the Department of Education.
–Agentic AI Supervision and Kill-Switch Compliance. Districts deploying any autonomous/agentic AI must ensure direct adult supervision, maintain complete activity logs, and implement mechanisms that allow school personnel to immediately suspend or disable the tool’s autonomous functionality without vendor assistance, all of which require dedicated staffing and technical infrastructure.
–Incident Reporting. The new requirement to disclose unexpected autonomous behavior or IT safety or security incidents involving approved AI tools, including unauthorized data access, adds another compliance layer requiring incident-detection, escalation, and reporting procedures.
–Legal Exposure. Noncompliance could expose districts to administrative enforcement, parental complaints, and litigation regarding vendor contract compliance under Section 1006.1494, Florida Statutes.
–Practical Implementation Tips Despite these challenges, school board employees should take these steps now:
–Begin Drafting Policy Amendments. Work with board counsel to draft amendments to your existing internet safety policy now, building in flexibility to adjust as the Department issues guidance.
–Conduct an AI Tool Inventory. Direct your IT department to compile a comprehensive inventory of instructional software, flag tools that incorporate AI or machine learning, and create an approval workflow with instructional, IT, and legal review. –
–Develop Training Protocols. Design tiered training: basic awareness for all employees, detailed compliance training for teachers, and specialized training for IT and administrative staff responsible for tool approval.
–Create Parent Notification Templates. Develop standardized, plain-language notification templates covering the tool name, applicable grades and subjects, nature of AI interaction, and opt-in instructions. Consider translating the templates into your district’s primary languages.
–Manage the Overbroad Definition Strategically. In your local policy, focus compliance efforts on AI instructional tools using generative or autonomous/agentic AI, the categories carrying the rule’s most significant obligations, rather than classifying every piece of software under the broad “AI system” definition. Document your interpretive rationale to demonstrate good-faith compliance.
–Evaluate Agentic AI Tools. Assess whether any currently deployed or proposed AI instructional tools qualify as autonomous/agentic AI under the new definition and, if so, ensure that adult supervision protocols, activity-logging systems, and kill-switch mechanisms are in place before the July 1, 2027 deadline.
–Establish Incident Reporting Procedures. Develop internal protocols for identifying and disclosing unexpected autonomous behavior or IT safety or security incidents involving approved AI tools to the Department as required by the amended rule.

The July 1, 2027 deadline may feel distant, but the scope of work is substantial. School board employees who begin planning now will be best positioned to achieve compliance while minimizing disruption and managing the significant, if officially unacknowledged, costs this rule will impose. AI is not the future; it is the present.