As states continue to develop guidance for districts around the use of artificial intelligence (AI) in K-12 education, it’s important to consider the role states play in supporting districts’ procurement of safe and effective AI tools. Currently, procurement policies related specifically to AI are primarily handled at the district level. Many district leaders have expressed a desire for more state-level guidance as they work to identify tools and weigh their costs and benefits.
States are in different stages of providing such guidance and support. Many states have existing software procurement policies that may encompass AI tool purchasing. However, these broader policies often lack provisions that would address concerns specific to the use of AI tools in education, including the use of student data to train an AI model, human-in-the-loop oversight or review requirements, bias and equity auditing, and end-user transparency or disclosure requirements, among other things.
State Examples
A few states have filled these gaps by requiring state agencies to develop model policies or guidelines regarding AI purchasing for school districts. Earlier this year, Idaho’s legislature passed S.B. 1227, which requires that procurement of any generative AI-related tools and applications complies with state and federal data privacy laws and allows the state department of education to establish a list of approved AI tools or model procurement guidelines. Maryland’s S.B. 720, also enacted earlier this year, requires the state department of education to develop a rubric to assist local school systems in evaluating AI tools.
Similarly, Alabama’s state department of education released a model AI policy in 2024 that includes an AI procurement section. The model policy requires a local education agency (LEA) to conduct a detailed needs and capability assessment before procuring an AI tool. In addition, the policy addresses AI-specific concerns by requiring an LEA to obtain a written statement from a provider that the AI model in question has been pretrained and that no data will be used to train a model to be used in the development of a new product. The written statement must also confirm that the provider has been responsible in the development of the AI system by using human-in-the-loop during model development and taking steps to minimize bias in the data selection process and algorithm development.
Utah has taken a different approach to AI procurement that simplifies district-level procurement concerns by implementing consortium purchasing for school districts across the state. Rather than providing guidance or model criteria to support school districts in identifying AI tools, the state selected and subsidized district access to certain AI tools following an RFP process. As a result, the state’s public schools were able to access three state-approved tools at low cost: SchoolAI, MagicSchool, and Skill Struck.
Several states are also currently considering legislation that may indirectly impact districts’ procurement of AI tools. For example, California A.B. 1159, if enacted, would expanding existing student data privacy laws to include protections against the use of student data for AI training. Districts looking to purchase AI tools would then be required to include such protections in any AI vendor’s contract. However, similar bills in other states failed to progress in the most recent legislative session for a host of possible reasons, including the uncertainty wrought by rapid and ongoing change in the field of AI and misalignment between proposed state regulations and federal AI priorities.
Nonetheless, understanding best practices for vetting and selecting AI tools for use in K-12 education remains a major concern for many state officials. As state policymakers weigh what kind of support school districts need, resources like the Southern Regional Education Board’s AI tool procurement, implementation, and evaluation checklist and the State Educational Technology Directors Association’s procurement guide may provide insight into additional questions and considerations policymakers should account for to maximize effectiveness and safety of AI tools in schools.
Disclosure: MagicSchool AI is a 2026 Education Commission of the States Event Sponsor.
