Conference Agenda

As artificial intelligence rapidly enters European classrooms, policymakers and researchers have mobilized to define what "AI literacy" means for teachers. International frameworks—UNESCO's AI Competency Framework, DigCompEdu 2.2, AI-TPACK models—promise to guide teacher education toward necessary competencies (UNESCO, 2024; Cosgrove & Cachia, 2025; Hu et al., 2024). While some frameworks acknowledge contextual variation, prevailing approaches to AI literacy often assume that competencies can be universally defined, measured, and standardized across different educational contexts. This panel challenges that assumption through empirical examination across seven European countries - Germany, Finland, Italy, Portugal, Austria, Lithuania, and Sweden - of how national contexts fundamentally shape what teachers need to know, do, and decide when working with AI.

Drawing on comparative research within the Erasmus+ Teacher Academy AI2PI (Project number: 101196121 – AI2PI), now in its second year, we explore how "AI literacy" is not a stable construct that can be transmitted uniformly but rather a contextually embedded practice shaped by infrastructure, policy, pedagogy, and culture. AI for lesson planning may mean working with institutional support and clear guidelines in one context, or improvising with limited tools while navigating restrictions and policy ambiguity in another. Frameworks apply the same competency descriptors to these incomparable realities.

Current frameworks often position context as an implementation factor—something to account for when rolling out pre-defined competencies. Yet our analysis reveals that context is constitutive, not incidental. What counts as necessary AI literacy depends on pedagogical autonomy, technological infrastructure, resource availability, teacher preparation quality, and policy support—all of which vary dramatically across European contexts.

These differences represent fundamentally distinct professional realities requiring varied forms of knowledge, judgment, and practice. A teacher navigating resource constraints, strict curricula, or policy ambiguity develops different competencies than one with abundant tools, pedagogical autonomy, and institutional support. How, as teacher educators, can we address AI literacy frameworks amid diverse professional realities?

Existing frameworks reflect their origins in CS/EdTech fields, emphasizing system-facing knowledge (how AI works, what it affords) while neglecting pedagogy-facing concerns (how it transforms teaching, relationships, values) (Sperling et al., 2024). This technical orientation privileges contexts with robust infrastructure, English dominance, and resources to experiment, creating standards that might measure access rather than competence, mistaking resource inequality for professional inadequacy (Adams et al., 2023; Henriksen et al., 2025). However, European contexts are characterised by pedagogical diversity: teachers across participating countries navigate varied infrastructures, multilingual classrooms, different policy environments, and distinct pedagogical traditions. Rather than treating this diversity as deviation from an ideal, we argue that it represents forms of professional expertise that frameworks should capture and validate.

This panel debates whether a unified European AI literacy framework is possible, and how context-responsive approaches might be integrated into teacher education. Can we design frameworks that honor contextual diversity rather than treating it as deviation from an idealized norm? The panel explores how to position teachers as authorities who define AI literacy within their contexts, rather than as recipients of externally-defined competencies. Panelists will surface tensions between standardization and contextualization that emerged during the first project year, explore how different national realities shape teacher needs, and consider what "unity in diversity" might actually require. The goal is to use these tensions as generative sites for reimagining how teacher education can adequately address AI's entry into classrooms.