The AI Literacy Heptagon: A Structured Approach to AI Literacy in Higher Education

Synthesis: Hackl, Müller, and Sailer (2026) present the AI Literacy Heptagon, a structured seven-dimensional framework for AI literacy (AIL) in higher education, developed through an integrative literature review of publications from 2021–2024. The framework synthesizes seven core dimensions — technical knowledge and skills, application proficiency, critical thinking ability, ethical awareness and reasoning, social impact understanding, integration skills, and legal and regulatory knowledge — and is operationalized through four Bloom's-taxonomy-aligned proficiency levels (Unaware, Beginner, Intermediate, Expert). An initial expert-led curriculum mapping across an AI Engineering and a Media Pedagogies and Teaching Strategies program demonstrated the framework's utility for analyzing and structuring curricula, highlighting the discipline-specific nature of AIL implementation.

Key Findings

• Seven dimensions emerged from an iterative coding of AIL conceptualizations: Technical Knowledge and Skills (TKS, 24/27 sources), Application Proficiency (AP, 26/27), Critical Thinking Ability (CTA, 23/27), Ethical Awareness and Reasoning (EAR, 20/27), Social Impact Understanding (SIU, 11/27), Integration Skills (IS, 18/27), and Legal and Regulatory Knowledge (LRK, 2/27).
• The authors deliberately retained emerging dimensions that appear infrequently in the literature — particularly Legal and Regulatory Knowledge (only 2/27 sources) and Integration Skills — arguing these address rapidly evolving regulatory landscapes (e.g., the EU AI Act) and the gap between theoretical knowledge and practical application.
• A synthesized working definition operationalizes AIL as critical, ethical, and responsible engagement with AI across all seven dimensions, explicitly acknowledging that emphasis varies by disciplinary context.
• The framework distinguishes generic AIL (Beginner level in all dimensions, the baseline for all students) from domain-specific extensions (Intermediate/Expert levels tailored to fields of study), with four proficiency levels mapped to Bloom's cognitive processes.
• Expert-led curriculum mapping of an AI Engineering and a Media Pedagogy program showed technical programs emphasize TKS and AP, while humanities-oriented programs emphasize EAR and SIU, supporting the framework's flexibility while maintaining multidimensional integrity.

Study Design & Method

The study used an integrative literature review (distinct from a meta-analysis, as it performs qualitative synthesis rather than statistical pooling), following PRISMA principles for transparency. A systematic search of Web of Science and Scopus for English-language publications from 2021 (plus earlier foundational works) to 2024 used term combinations around AIL, higher education, teaching/learning AI, and stakeholders. The search was completed December 10, 2024, with additional sources via citation tracking, Google Scholar, and research discovery tools. Two independent coders applied open then axial coding to extract recurring competencies, resolving discrepancies through discussion, yielding the seven dimensions. Initial framework validation used structured expert-led "collaborative profiling sessions" with two program leaders (AI Engineering and Media Pedagogy) to map curricula onto the heptagon and its proficiency levels — explicitly framed as evaluating curricular structure and goals, not empirically measuring student competencies.

What this means for practice

• Instructors. Audit your course against all seven Heptagon dimensions rather than equating AI Literacy with tool proficiency, and record which dimensions your modules currently leave untouched.
• Set Beginner-level competence in every dimension as the baseline for all students, and reserve Intermediate and Expert targets for the domain-specific extensions your discipline actually requires.
• Give legal and regulatory knowledge explicit curricular time: it was the most underrepresented dimension in the reviewed literature yet covers obligations such as the EU AI Act that graduates meet as AI deployers.
• Use the four proficiency levels (Unaware, Beginner, Intermediate, Expert) to run a competency-gap analysis of existing modules, then redesign the modules whose outcomes cluster in a single dimension.
• Design Assessment tasks that require students to combine several dimensions at once — for example, by having them justify a tool choice on technical, ethical, and legal grounds together — rather than testing technical skill and ethical reasoning as separate items — and make explicit where AI literacy ends and computational or data literacy begins.

Limitations

• The literature review is bounded to 2021–2024 English-language publications in Web of Science and Scopus, so relevant work outside that window or language may be missing.
• The initial validation is qualitative and small-scale — two expert-led curriculum mappings — explicitly framed as illustrating the framework's utility rather than empirically validating its effectiveness or measuring student competencies.
• The authors note the Visualization lacks granularity at the expert level, and that the knowledge-skills-attitudes three-dimensional nature is not explicitly represented in the visual model.
• The framework is a proposed instrument that still requires validation across more programs and disciplines.

Citation

Hackl, V., Müller, A. E., & Sailer, M. (2026). The AI literacy heptagon: A structured approach to AI literacy in higher education.