A critical review of teaching and learning artificial intelligence (AI) literacy: Developing an intelligence-based AI literacy framework for primary school education

University research (Faculty of Education context)

Cambridge, United Kingdom

Doctoral researcher with independent coding verification (60% of corpus)

• In-school (K–12)

Scoping review of peer-reviewed literature (PRISMA-informed) with thematic synthesis

Corpus search current to end of March 2024; multi-stage screening and coding

19 included studies representing 17 distinct frameworks

Not applicable (literature synthesis); reviewed frameworks reference classrooms, tools, and platforms

• Primary-specific AI literacy frameworks remain sparse relative to secondary and higher education
• Many existing frameworks lean on constructionism and technical artifact creation
• Cross-database search may omit grey literature or non-English sources
• Proposed framework requires future empirical classroom validation

Primary school students (explicit focus); corpus spans early childhood through university

Increasing informal exposure to AI assistants and media; formal understanding often limited

Not assumed for primary audience; paper critiques over-reliance on coding-centric pathways

• Map how AI literacy frameworks are produced and which educational levels they target
• Characterize methodological and theoretical orientations (e.g., Bloom-like progressions, constructionism, computation theory)
• Position AI literacy at the intersection of digital, data, and computational literacies with AI ethics
• Propose an inclusive primary-oriented framework emphasizing human–AI cognitive partnership (“AI thinking”)

• Support policymakers and curriculum designers with a holistic, transdisciplinary orientation
• Encourage ethical reasoning through interdisciplinary lenses (philosophy, economics, law, ecology)
• Promote equitable access by aligning expectations with primary cognitive development

• Not a randomized controlled trial of a classroom intervention
• Not a student assessment validation study of the new framework
• Not an exhaustive review of every database worldwide

Conceptual synthesis of AI literacy education (not a single commercial product)

• Co-creator

English-language peer-reviewed literature

• Frameworks reviewed describe varied learner interactions (unplugged activities, block-based tools, ML demos, GenAI use)
• Proposed “AI thinking” emphasizes mediated cognition through collaboration with AI as sociotechnical actor
• Ethical engagement includes scrutinizing data curation, bias, and societal impacts beyond narrow CS skill drills

• Treat AI outputs and training data as socially situated rather than neutral
• Integrate fact-checking, bias awareness, and accountability across subjects
• Design age-appropriate transparency about what AI can and cannot reliably do
• Avoid reducing AI literacy to uncritical tool consumption or shortcut completion

• Define inclusion criteria and search strings across four major citation databases
• Screen titles/abstracts, remove duplicates, and apply snowball sampling to linked frameworks
• Chart frameworks by level, methodology, theoretical orientation, themes, and learning foci
• Synthesize limitations of anthropocentric/constructionist dominance and derive design principles
• Present an intelligence-based framework figure and theoretical rationale (Vygotsky; post-humanism)

• Researchers lead corpus definition, coding, and interpretation with inter-coder verification on 60% of data
• AI systems appear as objects of study and as actors in ethical scenarios, not as authors of the review
• Educators remain responsible for translating framework principles into local curricula and safeguards

• Use developmental perspectives to sequence expectations for young learners
• Blend technical insight with creative and ethical inquiry across disciplines
• Document exemplar frameworks (e.g., Five Big Ideas, UMC progression, ABCE) as concrete anchors

• Literature reports gaps in assessment and outcomes evidence for very young learners
• Tension between hands-on constructionist activities and broader sociotechnical literacy needs
• Risk of technological determinism if AI progress is framed as inevitable rather than socially shaped
• Primary teachers may lack frameworks that match cognitive readiness without diluting ethics

• Shift from purely human-centric constructivism toward post-humanist co-agency of human and non-human actors
• Introduce “AI thinking” to name hybrid cognitive practices emerging from human–AI collaboration
• Recommend interdisciplinary ethics pedagogy rather than siloed “computer ethics” add-ons

• Identified 17 frameworks with diverse geographic origins (e.g., Hong Kong, United States, Germany)
• Documented rising publication momentum, especially in 2023, across educational levels

• Dominant development approaches include empirical studies, literature reviews, and standards alignment
• Recurring competency themes align with Long and Magerko-style dimensions: digital/data/CT/ethics and transdisciplinary skills
• Synthesis supports moving young learners toward responsible participation, not only technical consumption

The review concludes that holistic primary AI literacy should explicitly nurture cognitive, creative, and ethical engagement with AI as a societal actor, while acknowledging limitations (database coverage; partial independent coding) and calling for empirical tests of the proposed framework.

• Human-participant ethics approval was not required because the study is a literature scoping review
• Framework design emphasizes data justice, bias, and inclusive representation called out in prior harms literature
• Classroom use of any AI tool still requires local privacy policies, consent, and age-appropriate data minimization
• Authors declare no competing interests; open access licensing supports transparent reuse

• Activity documentation
• Practitioner observation

• Instrument design for measuring “AI thinking” constructs in primary grades
• Comparative curriculum studies pitting constructionist pathways against sociotechnical literacy pathways
• Longitudinal work linking early AI literacy framing to later critical evaluation competencies

• AI literacy conceptualization and framework evaluation
• Primary and early childhood computing education
• Post-humanist and sociocultural learning theory in EdTech

• Completed

Primary (K–5 focus in proposal)

Research synthesis → school implementation guidance

Literacy framing / ethics / transdisciplinary inquiry

Sociocultural mediation + interdisciplinary ethics

Low (synthesis); Medium once tools deployed in classrooms

Low for the review itself; variable when schools adopt tools