Utah has become one of the first states to codify comprehensive rules for artificial intelligence in K-12 classrooms. HB 273, the Classroom Technology Amendments, was signed into law on March 18, 2026, as Chapter 120 of the 2026 General Session, with an effective date of July 1, 2026 for its core provisions. The law requires every local education agency in the state to adopt a local AI use policy based on a model policy the Utah State Board of Education must publish by December 1, 2026. Local policies must be in place before July 1, 2027.
What the law requires
Under Utah Code § 53G-7-1402, generative AI may not independently grade student work or make high-stakes determinations, defined as decisions about placement, discipline, academic progression, or eligibility for programs and services. Students are barred from using generative AI to complete academic work unless a teacher authorizes it for a specific instructional purpose. When a school or educator uses generative AI for instruction, assessment, or classroom activities, the statute requires written parent notification that includes a link to the local education agency's metadata dictionary.
The law also prohibits AI tools from being used for biometric surveillance or psychological profiling. Approved AI tools must not expose students to content that is harmful, manipulative, or inappropriate. Local education agencies must maintain an inventory of approved AI tools authorized for educator or student use. After adopting a policy, the agency must submit it to the state board for review and approval, provide training to educators and staff, ensure compliance at each school, review the policy at least once every two years, and update it as needed. If an agency fails to adopt a policy or its policy is not approved, the USBE model policy becomes the default.
The implementing rule and draft model policy
The USBE has released a draft model policy, titled "Instructional Tech and AI Model Policy - DRAFT-V1- Aug 2026," and a draft implementing rule, R277-334 (Draft 1, dated May 27, 2026), both now circulating for public review. The model policy specifies nine required elements, including definitions, grade-level technology limits, parent transparency measures, AI standards, a resource plan for students with technology-related learning difficulties, AI sandbox course rules, and staff training requirements. Local education agencies may adopt the model policy as written or an amended version consistent with the statute and Rule R277-334.
Under Rule R277-334, before adopting a policy, a local education agency governing board must hold a public meeting allowing public comment under the Open and Public Meetings Act. After adoption, the agency must submit the finalized policy and a statement confirming the public meeting was held to the Superintendent. Agencies must annually certify compliance. If an agency fails to comply, the Board may require corrective action, additional reporting, additional training, or withhold or delay distribution of funds authorized under Section 53G-7-228. An agency receiving state digital literacy, computer science, or educational technology funds must also certify compliance as a condition of receiving those funds.
The AI sandbox course
The statute authorizes, but does not require, local education agencies to offer an AI Sandbox course for students in grades 9 through 12. The course would provide supervised instructional opportunities aligned with workforce development needs in emerging fields including AI, data science, and digital design. The course requires parental opt-in.
How other states compare
In 2026, 12 bills addressing classroom screen time were introduced across U.S. states. Approaches vary. Kansas proposed complete prohibitions on digital devices, while Iowa and Oklahoma proposed 60-minute daily caps for grades K-5. Utah and Alabama are the first to enact such laws. Alabama HB 78, signed March 4, 2026, requires the State Department of Human Resources and the State Department of Education to develop screen-time standards for early childhood education and training on those standards for teachers and staff, with a focus on pre-K students and limiting screen time in child care settings. Utah's law is more comprehensive in scope, covering both screen time and AI use across K-12.
Utah previously published an AI Framework for P-12 Education as guidance, not a binding policy, addressing AI use in schools, including compliance with data protection and privacy laws and transparency when users interact with AI versus humans. That framework predates the statutory requirements of HB 273.
What the research says
The legislative summary for HB 273 states that grade-level guardrails were "designed to match the research on student outcomes" and that the K-3 screen-time prohibition came directly from early elementary teachers reporting that screens were making behavioral issues worse. However, the document does not cite specific studies or evaluations; it references teacher testimony as the basis for the provision. The research base cited in the brief does not include any independent peer-reviewed studies or evaluations testing the effectiveness of similar policies.