Abstract:

The rapid expansion of Artificial Intelligence (AI) in society has highlighted the need for educational systems to equip students with the knowledge, skills and attitudes required to engage critically and responsibly with AI. Although several frameworks define the core dimensions of AI literacy, the assessment of these competences remains a major gap in both research and practice. Existing evaluation approaches ‒ mainly qualitative, fragmented and focused on technical understanding ‒ do not adequately capture the interplay between conceptual knowledge, informed use of AI systems and the ethical-critical dimension, as shown by literature.

Findings indicate that current AI literacy initiatives privilege operational and process-based knowledge while ethical reasoning, agency and critical thinking remain under-assessed. Moreover, learning activities often rely on active methodologies, yet validated competence-based assessment tools are largely absent. To address this gap, this contribution proposes a competence-based assessment grid for AI literacy grounded in the AI-related indicators of DigComp 2.2. The framework integrates the three foundational dimensions of AI literacy ‒ Knowledge; Skills and Attitudes; Ethics and Critical Thinking and aligns them with the Engage-Investigate-Act structure of Challenge-Based Learning, a pedagogical model designed to support authentic, performance-based evidence.

The resulting grid is organised into five areas: understanding what AI systems can and cannot do; recognising how they work; interacting with AI tools consciously; verifying outputs; identifying ethical and societal risks; and exercising human agency. For each area, the grid provides specific indicators, progressive levels of achievement and examples of authentic evidence including scenario-based reasoning, prompt iteration, bias identification, evaluation of AI-generated content and decision-making tasks. The integration of Challenge-Based Learning ensures that assessment focuses on observable behaviours and real-world performances rather than declarative knowledge alone.

The proposed framework offers teachers and curriculum designers a transparent, systematic and pedagogically grounded tool for assessing AI literacy. It bridges the gap between competence definitions, active learning practices and classroom evaluation, and lays the foundation for future empirical validation studies.

Keywords:

AI literacy, Assessment, K-12 Education, Challenge-Based Learning, Ethics, Critical thinking.