As generative artificial intelligence becomes integrated into educational systems, governments are producing guidance documents to manage its use in K-12 schools. While these documents are often framed as technical or procedural resources, they also shape normative ideas about how AI should be perceived, trusted, limited, and incorporated into learning settings. This research views national and regional AI guidelines as cultural texts that reflect emerging social and technical visions of education in the age of AI.
This study presents a comparative thematic analysis of K-12 AI guidance documents from the United States and the United Kingdom. The United States operates within a federal system in which individual states exercise important authority over K-12 education policy, resulting in heterogeneous AI guidance frameworks across jurisdictions. By contrast, the United Kingdom, as a unitary state with devolved educational authority, distributes responsibility for schooling across England, Scotland, Wales, and Northern Ireland, creating nationally coordinated yet regionally differentiated policy frameworks. By comparing these two governance models, this project examines how AI education and instruction are designed, implemented, and pedagogically framed within K-12 education systems in different countries. It analyzes how these subjects are prioritized, how risks and responsibilities are articulated, and what differing approaches arise in their direction across national contexts.
The dataset consists of publicly available state-level AI guidance documents in the United States and national or regional guidance issued by U.K. educational authorities. Using iterative qualitative coding, the study identifies recurring thematic clusters in both contexts. Although AI is viewed as both an opportunity and a risk in both countries, the 2023-2024 U.S. state-level AI guidance documents indicate an initial phase of stabilization in K-12 AI governance. This phase is characterized by academic integrity concerns, human-centered boundary setting, equity rhetoric, and gradual movement toward organized implementation frameworks. Notably, across both national contexts, equity and access concerns appear nearly universal; however, in several U.S. states in particular, academic integrity anxiety emerges as a primary early driver shaping the trajectory of AI policy discourse. In addition, in the U.S. state guidance documents, nearly every state includes some articulation of AI literacy, typically emphasizing understanding how AI systems work and fostering critical evaluation of AI-generated outputs. However, the concept of AI literacy is defined inconsistently across states. Some states conceptualize it as technical proficiency, others as ethical discernment or digital citizenship, and a few embed it within existing computer science standards. This divergence reveals a broader conceptual ambiguity in how AI literacy is constructed and institutionalized within K-12 policy discourse.
Preliminary results reveal several differences in emphasis across different national contexts. Overall, U.S. state documents often present AI in terms of innovation, workforce readiness, and institutional responsibility, frequently highlighting concerns about academic integrity and plagiarism. U.K. guidelines, on the other hand, more consistently place AI within frameworks of safeguarding, child protection, and data governance influenced by broader regulations like GDPR. Although both countries emphasize AI literacy, their underlying concepts differ. In the U.S., literacy is often seen as an adaptable technological skill, whereas in the U.K., it is more closely linked to digital citizenship and responsible use. Furthermore, there is variation in the depth of policy operationalization. Certain states provide detailed assessment rubrics, phased integration models, and implementation checklists, whereas others articulate principles at a rhetorical level without translating them into actionable frameworks. Lastly, it is notable that the policy frameworks referenced differ across national contexts. In many cases, AI governance is layered onto existing privacy and data protection structures rather than being developed as a standalone regulatory domain.
Keywords: K-12 AI Education, AI Literacy and Ethics, Comparative Analysis