From AI intuition to AI literacy: A dual framework for K-12 education

Synthesis: This position paper (Beau & Lazar, EdArXiv 2026) argues that existing AI Literacy frameworks — with their emphasis on conceptual understanding, practical application, critical evaluation, and ethical judgment — implicitly assume learners approach AI through declarative, rule-based knowledge. Yet in practice, K-12 students (and teachers) encounter generative AI far more experientially: experimenting with prompts, observing system behavior, and adapting strategies before they can articulate formal principles. To account for this, the paper introduces AI intuition — an experiential, inductive, often tacit form of understanding developed through iterative interaction with AI systems that supports context-sensitive judgment under uncertainty — as a complementary construct to AI literacy. It proposes a dual framework that situates AI literacy as a structured, largely static set of competencies alongside AI intuition as a dynamic learning process, grounded in classroom implementation at the International School of Boston and oriented toward the 2029 PISA Media and AI Literacy standards.

The under-theorized experiential dimension

The paper's premise is that K-12 AI-education research has focused on structured competency progressions (understanding how AI works, using it effectively and responsibly, critically evaluating outputs, awareness of ethical and societal implications) while largely under-theorizing how learners actually come to know AI. Students frequently develop a practical "feel" for how AI responds to different prompts, constraints, and contexts before they can articulate formal rules or technical explanations. The notion of intuition has a long history in philosophy, cognitive science, and education as knowledge that is experiential, inductive, and often tacit and that precedes or complements formal reasoning (Dewey, Hogarth, Kahneman). In STEM education, intuitive understanding has long been recognized as a necessary precursor to formalization, abstraction, and proof — yet intuition has rarely been discussed explicitly in AI education.

An exception is the recent work of Flechtner and colleagues, who introduced AI intuition in university-level design education to describe an experience-based understanding of AI developed through playful experimentation, embodied role-play, and hands-on prototyping. This paper extends that concept to K-12, where it has not previously been applied.

The dual framework

AI literacy and AI intuition are framed not as sequential stages but as interdependent dimensions of AI fluency:

• AI literacy — structured, reflective, ethical, and conceptual knowledge about AI systems. It provides conceptual clarity, ethical Guardrails, and critical frameworks.
• AI intuition — an experiential, inductive, and embodied understanding of how AI behaves in practice, developed through guided exploration, experimentation, and Reflection. It provides the practical wisdom to apply knowledge in dynamic, uncertain contexts.

The framework maps the two across five complementary learning dimensions:

1. Understand AI concepts — AI literacy gives structured knowledge of how AI works (data, algorithms, bias, learning); AI intuition gives a "gut feeling" that AI behavior is context-sensitive and sometimes unpredictable.
2. Use AI tools effectively and creatively — literacy means applying AI responsibly in tasks; intuition means experimenting with prompts, observing patterns, and refining strategies through trial and error.
3. Evaluate AI critically — literacy means verifying accuracy, assessing bias, and considering fairness/reliability; intuition means recognizing limits, errors, and surprises through direct observation — supporting a "trust but verify" stance.
4. Apply AI ethically and responsibly — literacy covers privacy, Academic Integrity, and societal impact; intuition means strategically deciding when AI is valuable and when to avoid it.
5. Reflect and strategize — literacy supports conceptual reflection (big picture); intuition supports practical reflection and the building of "rules of thumb" for daily use.

AI intuition as a safeguard against over-reliance

A core argument is that cultivating AI intuition serves as a safeguard against technological over-reliance and the erosion of critical thinking. By engaging in inductive experimentation and "phenomenological mapping" — for example stress-testing prompts to identify where a model fails — students move from passive consumers to "expert" observers of AI behavior. This hands-on, active experimentation develops a tactile understanding that lets learners recognize hallucinations and bias through direct, playful engagement. Importantly, the paper distinguishes AI intuition from prompt engineering: prompt engineering focuses primarily on optimizing outputs, whereas AI intuition emphasizes epistemic judgment — knowing when to trust, verify, constrain, or disengage from AI systems. The pedagogical path runs through experiential learning, dialogical moments where students articulate and revise their positions based on both experiential evidence and ethical considerations, thereby connecting personal experience with shared norms of responsible use and reinforcing AI literacy.

Directions for research and practice

The paper calls for: longitudinal studies of how cultivating AI intuition affects students' critical thinking and their ability to navigate evolving algorithmic societies; a more robust body of classroom-based research providing evidence-based instructional strategies (and assessment) for fostering intuitive "gut feelings" alongside technical literacy; specialized teacher professional development for facilitating inductive, exploratory learning environments; and domain-specific frameworks exploring how AI intuition manifests across arts, humanities, and advanced sciences. It also flags the strategic importance of the 2029 PISA Media and AI Literacy assessment for aligning K-12 AI education internationally, arguing that grounding technical learning in experiential knowledge is the best way to keep AI education ethical, inclusive, and sustainable.

Contribution to the knowledge base

This paper adds an experiential/epistemological dimension to the AI Literacy literature that most framework work under-specifies. Where the wiki's existing literacy pages catalog competencies, standards, and teacher competencies, Beau & Lazar supply a theory of how the competency develops — that learners acquire a working, inductive feel for AI before (and alongside) formal rules, and that instruction should deliberately cultivate both. It connects to Experiential Learning, constructivism (learning through direct engagement), and Critical Thinking, and it complements K-12 treatments of AI education with a concrete dual-process account of learner appropriation. It is a position paper, so its framework is conceptual awaiting empirical validation, but it grounds the claim in classroom implementation and gives AI-education research a usable construct ("AI intuition") for studying experiential learning with AI.

What this means for practice

• Instructors. Run every AI-literacy lesson through the paper's four-step cycle — exploration, observation, reflection, adaptation — so students build a working feel for model behavior instead of memorizing rules about it.
• Instructors. Assign phenomenological mapping: have students stress-test prompts to locate where a model fails, then articulate the pattern they found, which turns hands-on failure into explicit Critical Thinking about Hallucination Risk.
• Curriculum designers. Teach the five dual-framework dimensions in pairs, so declarative AI Literacy content (data, bias, ethics) always arrives with an Experiential Learning task that lets students induce the same principle from direct interaction.
• Researchers. Build assessment for AI intuition itself; the paper reports no evidence-based instruments or instructional strategies for the intuitive dimension, only that they are needed.

Limitations

• The paper is a position paper: it reports no empirical study, no sample, and no measured student outcomes, so the dual framework's claims about AI intuition remain conceptual.
• The classroom grounding is an implementation at one school (the International School of Boston), offered as context for the framework rather than as study data with participants and measures.
• The framework's stated validation target, the PISA 2029 Media and Artificial Intelligence Literacy assessment, has not yet been administered, so alignment with it cannot be checked.
• The authors themselves call for longitudinal impact studies and classroom-based research, conceding that evidence linking cultivated AI intuition to improved Critical Thinking does not yet exist.

Citation

Beau, M., & Lazar, M. (2026). From AI intuition to AI literacy: A dual framework for K-12 education. EdArXiv preprint.