Poster #135 - Teachers as AI Policy Actors: Focus Group Evidence from a K-12 Robotics Consortium

• In Event: Saturday Poster Luncheon

Abstract

Artificial intelligence tools are diffusing into K-12 STEM classrooms at a pace that outstrips corresponding federal guidance (Holmes et al., 2019, and Akgun & Greenhow, 2022), while state agencies, school districts, and nonprofit consortia shape the terms of classroom AI use in fact. Teachers operating under such policy ambiguity function as street-level bureaucrats whose discretionary judgments constitute a form of emergent policymaking (Lipsky, 2010). Their needs, constraints, and attributions of accountability constitute an under-examined input to AI-in-education policy design. This study asks: as AI tools are introduced into pre-college instruction, what needs, constraints, and attributions of accountability do K-12 teachers articulate?We conducted three focus groups with K-12 teachers who participated in a federally funded multi-state robotics curriculum in the 2025-2026 academic year. The first focus group, in April 2025, was devoted to an AI-tool wishlist. The next two, in June 2025, followed hands-on Python-curriculum training. Transcripts were analyzed through deductive thematic analysis (Braun & Clarke, 2006), applying a three-tier coding framework that operationalizes Lipsky's (2010) street-level bureaucracy concept as (i) teacher-articulated demand, (ii) implementation constraint, and (iii) accountability attribution.Teachers' AI-related demands concentrated on in-the-moment instructional assistance and curricular integration. Professional development, which teachers raised most frequently in non-AI discussions, was mentioned far less often once the conversation shifted to AI. Teachers frame AI primarily as support for their own technical work, such as code debugging, lesson design, and grading, rather than as a differentiation tool for addressing student heterogeneity. In non-AI discussions, teachers' most prominent constraint was time and capacity. In AI discussions, policy and mandate ambiguity, together with tensions around pedagogical philosophy, became the most frequent constraints. Equity-related constraints remained stable. Budget did not figure centrally in AI discussions. Contextual differences are most pronounced in accountability attribution. Teachers rarely named industry in non-AI discussions. In AI discussions, industry emerged as the most frequently mentioned actor, and school districts were named substantially more often. Federal and state actors appeared only sporadically.The findings suggest three implications for state-level AI education policy. First, teachers' core perceived AI constraint is policy-legitimacy ambiguity. States and districts should prioritize clear, actionable classroom AI-use guidelines. Second, teachers' conceptions of AI have not yet connected to differentiated instruction, leaving room for equity-sensitive tool design for English learners and students with limited home connectivity. Third, industry has de facto become the most frequently identified actor for AI implementation, yet this role lacks an accountability framework. State and consortium regulators should develop arrangements that recognize industry-consortium collaboration as the dominant implementation mode.

Authors

• Jialing Wu, The Ohio State University
  Presenting Author

• Daniel Li, Ohio State University
  Non-Presenting Co-Author