AI Literacy Heptagon Framework

• The AI Literacy Heptagon is a competency framework that defines AI literacy through seven interdependent domains covering technical, ethical, social, and legal aspects.
• It employs spider-web profiles and staged proficiency levels to systematically map, assess, and refine curricula in both academic and professional settings.
• By integrating empirical models and interdisciplinary pedagogical strategies, the framework offers actionable insights for enhancing AI literacy education.

The AI Literacy Heptagon is a multidimensional competency framework synthesizing contemporary definitions, empirical models, and curriculum design strategies for AI literacy, particularly within higher education and professional settings. It conceptualizes AI literacy (AIL) as encompassing seven equally weighted, interdependent domains. Each vertex of the heptagon corresponds to a dimension critical for engaging, evaluating, and integrating AI—technically, ethically, socially, and legally. The heptagonal structure not only serves as a pedagogical model but also scaffolds the curriculum mapping, assessment, and ongoing development of AI literate graduates and professionals (Hackl et al., 23 Sep 2025, Kennedy et al., 26 Oct 2025, Carolus et al., 2023).

1. Definitions and Theoretical Foundation

AI literacy is defined as the set of competencies enabling individuals to critically evaluate AI technologies, communicate and collaborate effectively with AI systems, and deploy AI as a tool across diverse contexts online, at home, and in the workplace (Long 2020). Within higher education, AIL is recognized as an essential 21st-century skill, reinforced by policy mandates (e.g., EU AI Act) and underlined by the need for systematic, comprehensive curricular integration (Hackl et al., 23 Sep 2025).

The AI Literacy Heptagon was created following integrative literature reviews and psychometric modeling (notably by Hackl et al. and the AI & Data Acumen Framework) to overcome fragmented approaches to AIL in educational practice and to bridge competing theoretical definitions. The model draws from foundational work in AI literacy taxonomy (Ng et al.), critical thinking (Lai 2011), self-efficacy (Bandura), and ethical AI scholarship (Hackl et al., 23 Sep 2025, Kennedy et al., 26 Oct 2025, Carolus et al., 2023).

2. The Seven Dimensions of AI Literacy

The core of the heptagon consists of seven mutually reinforcing dimensions. The specific facets—along with typical operationalizations and examples—are summarized in the following table for major frameworks:

Dimension	Definition (Hackl et al.)	Example Indicators (Kennedy & Gupta)
Technical Knowledge	Fundamental AI principles, algorithms, models, their operation and limitations	Data preprocessing, model selection, security, type I/II error trade-offs
Application Proficiency	Ability to use AI tools effectively across contexts	Effective use of AI writing assistants, coding aids
Critical Thinking	Analytical, evaluative reasoning about AI systems and their outputs	Framing questions for AI, interpreting outputs, addressing algorithmic bias
Ethical Awareness	Engaging with individual, societal, environmental ethical questions surrounding AI	Bias detection, privacy safeguards, compliance with norms
Social Impact	Understanding AI’s long-term economic, political, societal effects	Digital equity, political implications, community context
Integration Skills	Embedding AI in digital/organizational workflows, human-AI collaboration	Human-AI co-design, workflow redesign, participatory integration
Legal/Regulatory Knowledge	Understanding laws (EU AI Act, GDPR, copyright, competition law) affecting AI deployment	Compliance, legal risk assessment, regulatory differentiation

Alternative models include “Self-Efficacy,” “Collaboration,” “Innovation & Creativity,” and “Cognitive” as heptagon facets. The MAILS scale empirically isolates Use & Apply AI, Understand AI, Detect AI, AI Ethics, Create AI, AI Self-efficacy, and AI Self-management, with distinct psychometric support (Kennedy et al., 26 Oct 2025, Carolus et al., 2023).

3. Structural Properties and Interrelations

The heptagonal geometry is pedagogically motivated: its symmetry conveys the equal importance of all dimensions, demanding balanced development. No single vertex dominates; technical mastery without ethical and social competencies may result in misuse, while legal compliance without organizational integration yields fragile implementations. The seven-point structure is cognitively tractable for curricular design and richer than lower-dimensional constructs (Hackl et al., 23 Sep 2025).

Mutual reinforcement among dimensions has theoretical and empirical grounding. For instance, self-efficacy and self-management are found to strongly correlate with core AI literacy ($r = .72$–$.83$), supporting the theory of planned behavior. “Create AI” is psychometrically separate yet moderately correlated ($r \approx .50$) with other literacies, pointing to the need for explicit instruction and assessment rather than presuming transfer from user-level skills (Carolus et al., 2023).

4. Implementation in Pedagogy and Assessment

The AI Literacy Heptagon supports systematic curriculum mapping, aligning courses and learning outcomes to each dimension. Proficiency is typically staged on four levels—Unaware, Beginner (Remember, Understand), Intermediate (Apply, Analyze), and Expert (Evaluate, Create)—often adapted from Bloom’s taxonomy (Hackl et al., 23 Sep 2025, Kennedy et al., 26 Oct 2025). Assessment instruments (e.g., SNAIL, AILQ, AICOS, MAILS) use these dimensions to generate spider-web profiles for individuals or programs, revealing gaps and strengths.

A quantitative scoring rubric proposed by Kennedy & Gupta computes an overall AI literacy percentile:

$$\text{Score} = \frac{1}{7} \sum_{d=1}^{7} \frac{L_d}{4} \in [0,1]$$

where $L_d$ is the highest proficiency level demonstrated in dimension $d$.

5. Curriculum Case Studies and Educational Practice

Practical implementation is illustrated by expert-led mappings across domains. In an applied AI Engineering program, “Technical Knowledge and Skills” is achieved at intermediate level via core machine learning and deep learning courses; “Application Proficiency” and “Integration Skills” are met through project-based modules; “Legal and Regulatory Knowledge” is typically addressed at the beginner level in dedicated law-and-AI seminars (Hackl et al., 23 Sep 2025).

In a Media Pedagogy curriculum, critical thinking (“Media Pedagogical Questions”), societal impact (“Media Didactics”), and foundational technical knowledge are mapped to beginner or intermediate levels, with targeted improvement areas identified via heptagon profile “gaps” (notably Legal/Regulatory Knowledge) (Hackl et al., 23 Sep 2025). The AI & Data Acumen framework further recommends integrating heptagon-aligned assignments—e.g., cross-disciplinary team hackathons (Collaboration, Technical), reflective e-portfolios (Self-Efficacy), and policy memos (Cognitive, Socio-Cultural) (Kennedy et al., 26 Oct 2025).

6. Assessment Instruments and Psychometric Models

The Meta AI Literacy Scale (MAILS) operationalizes the heptagon through 60 diagnostic items plus meta and psychological competencies, supported by confirmatory factor analysis. The factor structure demonstrates robustness (Cronbach’s $\alpha > 0.85$ for all facets) and supports diagnostic, developmental, and program-integration use cases (Carolus et al., 2023). Higher-order modeling distinguishes between “core AI literacy,” “AI self-efficacy,” and “AI self-management,” facilitating both granular diagnostics and comparative research.

Assessment practices informed by these instruments include formative quizzes on knowledge and ethics, peer-reviewed projects against ethical checklists, and capstone projects demanding technical, ethical, and integration mastery (Hackl et al., 23 Sep 2025, Carolus et al., 2023).

7. Challenges and Recommendations for Implementation

Implementation frequently encounters challenges such as resource constraints, disciplinary silos, and the rapid obsolescence of content due to AI’s pace of change. Best practices include:

• Starting with small-scale pilots embedding two or more dimensions before program-wide adoption.
• Forming interdisciplinary teaching teams for integrative delivery.
• Regular content versioning and subscription to policy/industry updates.
• Embedding socio-cultural and ethical reflection across technical assignments.
• Iterative program profiling via the heptagon and alignment with workforce or regulatory needs (Kennedy et al., 26 Oct 2025).

Maintaining coverage across all seven dimensions—at least at the “beginner” level—is recommended, with advanced domain-specific depth developed in higher program stages. Ongoing empirical validation (e.g., via Delphi studies or benchmarking) is also encouraged (Hackl et al., 23 Sep 2025).

---

The AI Literacy Heptagon thus constitutes a harmonized, evidence-based framework designed to ensure higher education and professional programs cultivate graduates with the full spectrum of competencies required for responsible, effective, and adaptive participation in AI-infused societies (Hackl et al., 23 Sep 2025, Kennedy et al., 26 Oct 2025, Carolus et al., 2023).