Experts tell Joint Interim Committees to pilot and measure AI in schools while strengthening safeguards for children

Co-chair Representative Robbins convened the Joint Interim Committees for a hearing on artificial intelligence and children that featured researchers, policy advocates and local practitioners.

Kevin Frazier, senior fellow at the Abundance Institute, told the committee that lawmakers should evaluate AI “compared to what,” and avoid simplistic bans. “AI is a tool,” Frazier said, and policy should focus on the tool’s capabilities, the user’s awareness, the environment where it is used and whether errors can be corrected. He recommended short, measured pilots—six to 18 months—followed by audits and sunset provisions so laws can be revised as the technology evolves. Frazier highlighted that only a small share of teens are constant AI users but cautioned that sensitive uses (he cited Pew Research figures) require tailored protections.

Rob Ellavel, co-founder of Transparency Coalition, pressed for a layered legislative approach that combines narrow bills, privacy protections and enforcement. He criticized legal carve-outs that he said can exempt major platforms and urged states to use product-liability and tort law to require testing and documentation of AI products used with minors. “They assume tech laws will not be enforced in The States,” Ellavel said, and he recommended avoiding cure periods that let companies change behavior after notice, making enforcement impractical.

Professor Arbel of the University of Alabama School of Law described categories of AI exposure—general assistants, companions, school AI, embedded AI and emerging “agentic” AI that plans and executes tasks—and recommended three regulatory levers: pre-deployment standards where appropriate, harm-based (post-deployment) remedies, and positive integration through procurement and training. He urged duties of care (what developers owe children), clear reporting duties when an AI signals serious harm, and funding for educator training so teachers can use AI without worsening “cognitive deskilling.”

Randy Sparkman of AI Huntsville described local pilots and workforce training, telling lawmakers that practical vocabulary and consistent definitions across bills help both regulation and procurement. He described a high-school AI course and partnerships with community colleges to build a pipeline for future jobs.

Committee members asked detailed questions about cognitive deskilling, job displacement, and the feasibility of age limits. Frazier and Arbel both argued that blanket bans are easy to evade and not always appropriate; instead, they favored layered defenses—age-aware design, parental controls, school network gating, and stronger reporting and procurement demands for vendors working with schools.

The committee directed Legislative Services (LSA) staff to research specific model bills and enforcement options and scheduled a follow-up meeting focused on data privacy. Presenters left contact information for follow-up questions and recommended drafting iterative, narrowly targeted bills with measurable pilot periods.

The committee’s next meeting will focus on data privacy; staff were asked to assemble relevant state examples and the bills referenced during the hearing.