AI Pre-Literacy

Priority and Scope of the Coinage

We introduce AI Pre-Literacy as a new developmental construct. The term is coined to designate a bounded territory of childhood experience that has emerged as a consequence of contemporary artificial intelligence and that existing literacy frameworks — emergent literacy, digital literacy, media literacy, AI literacy, computational thinking, algorithmic literacy, and data literacy — do not jointly encompass. This document constitutes the primary citable statement of the construct. It is a conceptual introduction, not an empirical validation. It articulates the formal definition, the philosophy behind the coinage, the boundary conditions distinguishing the construct from adjacent frameworks, the five pillars, the anticipated objections, the frequently asked questions, and the falsifiability conditions. Empirical work is in progress.

Development from the Initial Proposal

The term AI pre-literacy was first introduced in Pasha and Pasha (2026), which proposed four provisional dimensions — AI-related vocabulary, imaginative and cognitive sense-making, attitudes and perceptions, and emerging socio-ethical awareness — and explicitly declined to specify an age range. The present document is the expanded formal statement of the same construct. Three developments are deliberate. First, empathic attunement is added as a fifth pillar, on the ground that differentiating simulated from human affect is a distinct developmental task not captured by the original four dimensions. Second, a prototypical range of birth to eight years is now specified, following the onset of AI literacy frameworks at age eight; the range remains prototypical, not maturational. Third, the reading of pre as prime — the first, foremost, formative years — is now made explicit, to prevent the deficit interpretation that “pre” invites. The construct is unchanged in its core claim: AI Pre-Literacy concerns developmental foundations, whereas AI literacy concerns demonstrated competencies. What has changed is the articulation.

Core Philosophy

The construct rests on one premise: AI is now woven into the fabric of human society, and preparing children for that society requires a developmental approach distinct from the one inherited from the Industrial Revolution. The industrial model produced childhood as a preparatory period for a standardized workforce — sequential, measurable, receptive. That model is inadequate for a society in which AI mediates work, learning, relationships, and civic life.

This is why we take pre in the sense of prime. The 0–8 period is not the time before AI literacy; it is the prime formative territory in which a child’s orientation toward AI is laid down — whether as a consumer of systems designed elsewhere, or as a master of systems she can interrogate, direct, and hold to account. AI Pre-Literacy exists to name that territory so that it can be deliberately cultivated rather than left to commercial design.

This is a rights claim as much as a developmental one. The United Nations Convention on the Rights of the Child guarantees children’s voice (Article 12), access to media (Article 17), and protection from exploitation. General Comment 25 (2021) extends these rights explicitly to the digital environment. AI Pre-Literacy is the developmental foundation on which these rights become exercisable in an AI-woven society. A child who cannot interrogate, refuse, or hold AI to account cannot meaningfully exercise voice in the environment that now shapes her life.

AI Literacy asks what children should know about AI. AI Pre-Literacy asks what childhood has become, now that AI is woven into the fabric of society.

1. Formal Definition

AI Pre-Literacy is the developmental territory from birth to approximately eight years in which young children encounter, notice, represent, interpret, discuss, orient toward, and respond to artificial intelligence and AI-mediated phenomena — forming the linguistic, cognitive, affective, empathic, and socio-ethical foundations upon which later AI literacy, critical evaluation, and responsible participation may develop, within the dynamic social, cultural, linguistic, technological, familial, educational, and community ecologies of childhood.

The definition is deliberately parallel to, yet distinct from, emergent literacy (Clay, 1966; Teale & Sulzby, 1986; Whitehurst & Lonigan, 1998). Where emergent literacy describes the developmental foundations of print literacy, AI Pre-Literacy describes the developmental foundations of AI sense-making.

2. The Necessity of a Different Approach

2.0 A Difference of Approach, Not of Degree

AI Literacy and AI Pre-Literacy begin from opposite vantage points, and the difference is not one of age or content but of orientation.

This is why AI Pre-Literacy is not AI literacy for younger children. A simplified version of a competency framework remains a competency framework. Reducing the content, slowing the pace, and replacing text with pictures does not change the object-oriented logic. AI Pre-Literacy does not teach children about AI. It asks what childhood development now requires because AI is woven into the fabric of society.

The distinction has three consequences. First, AI Literacy presupposes a learner who can read, reflect, and reason abstractly; AI Pre-Literacy addresses the period before those capacities are available. Second, AI Literacy measures what a learner knows and can do; AI Pre-Literacy describes the orientations from which all later knowing and doing will proceed. Third, AI Literacy produces competent users of AI; AI Pre-Literacy aims at developmental sovereignty — children who author with AI rather than being authored by it.

The two are sequential, not competing. AI Pre-Literacy names the prime territory in which the foundations of AI Literacy are laid. But they must not be conflated, because conflation produces the error of treating early childhood as a delivery site for simplified technical content — an error that would reproduce, in the AI order, precisely the industrial-era mistake that emergent literacy was coined to correct.

2.1 The Structural Rupture

Childhood has never developed independently of its technological order. This claim is empirically grounded across four paradigm shifts. The industrial order produced modern childhood as a protected, schooled category (Ariès, 1962; Cunningham, 1991). The digital order re-described children as developing differently because they are surrounded by digital tools from birth (Prensky, 2001; Buckingham, 2007). The information order re-situated that development within networked structures that process information through microelectronics (Castells, 1996; Livingstone, 2009). The AI order now mediates not only tools but parenting, learning, and sense-making itself, creating a structural rupture in which children can produce outputs they cannot yet understand (Livingstone & Blum-Ross, 2020; UNESCO, 2021, 2023; Holmes et al., 2021).

The rupture is not merely incremental. A four-year-old can now produce a visually accomplished landscape through a few spoken words to a generative AI system — an output once reserved for trained painters after years of formation. The child has not gained the skill; the child has gained the output. The gap between what a child can produce and what a child understands has widened structurally, not merely quantitatively.

2.2 The Industrial Invention of Pre-Literacy

Pre-literacy in the Industrial Revolution was introduced as reading readiness to make mass schooling governable. Industrialism removed children from factories and redefined childhood as a protected, preparatory period (Ariès, 1962; Cunningham, 1991). To produce a standardized workforce, schools needed a measurable sequence: precursors before conventional literacy (Morphett & Washburne, 1931). That deficit model — implying children were “pre-anything” — was later overturned by Clay (1966) and Teale and Sulzby (1986), who coined emergent literacy to argue that children are from birth “in the process of becoming literate” (Teale & Sulzby, 1986, p. xix), formalized as the skills, knowledge, and attitudes that are developmental precursors to reading and writing (Whitehurst & Lonigan, 1998). Thus pre-literacy was invented to bridge the gap between the child as labourer and the child as learner.

Within literacy studies, early twentieth-century discourse relied on terms such as “pre-reading” and “reading readiness,” implying a period before literacy proper began. Emergent literacy is not “pre-anything”; there is no point in a child’s life when literacy begins. We make a parallel but distinct argument today. The historical lesson is that pre-literacy was not a natural developmental fact but an institutional invention — one that emergent literacy later reconstituted on non-deficit grounds. AI Pre-Literacy is proposed in the same spirit: not as a deficit category, but as a bounded territory of foundational sense-making that existing frameworks have left unnamed.

2.3 Why AI Pre-Literacy Is More Than Emergent Literacy

AI Pre-Literacy is more than emergent literacy because AI has transformed childhood from learning to decode a human-authored world (Teale & Sulzby, 1986) to learning to interrogate a world that simulates authorship itself (Turkle, 2011). Where emergent literacy bridged oral language to print (Whitehurst & Lonigan, 1998), AI now divorces capability from comprehension by allowing children to produce expert outputs before they can read (UNESCO, 2023; Holmes et al., 2021). Where print required literacy, AI requires foundations to calibrate trust, detect bias, and differentiate simulated affect (Long & Magerko, 2020; Livingstone & Blum-Ross, 2020). And where print was consumed, AI is relational — children attribute intentions and emotions to systems that have none (Kahn et al., 2012), making childhood in the AI era qualitatively more demanding than its predecessors (Su et al., 2023).

The distinction is nonetheless not a matter of degree. Emergent literacy and AI Pre-Literacy share a developmental logic — foundations before formal competence, non-deficit framing, ecological embeddedness — but they diverge in the phenomena they address. Emergent literacy concerns the transition from oral to written language in a human-authored symbolic order. AI Pre-Literacy concerns the transition from naïve to calibrated sense-making in a computational order that simulates authorship, intention, and affect. The developmental tasks are therefore different in kind: not phonological awareness and print concepts, but source attribution, epistemic trust calibration, and empathic differentiation.

2.4 The Gap in Existing Frameworks

AI literacy frameworks have responded to this rupture by defining competencies to know, use, evaluate, and create AI (Long & Magerko, 2020; Ng et al., 2021a, 2021b; Touretzky et al., 2019). These frameworks, however, assume a learner who can already read, reflect, and reason abstractly. Even when extended downward in policy guidance (UNESCO, 2021, 2023), they describe curricular content rather than developmental precursors. They do not describe how young children, before conventional literacy, first encounter, represent, and orient toward AI-mediated phenomena in the ecologies of home, preschool, and community. It is this pre-competency developmental territory — not merely new curriculum content — that AI Pre-Literacy exists to name, bound, and investigate.

2.5 Boundary Work: AI Pre-Literacy and Adjacent Constructs

The distinction between AI Literacy and AI Pre-Literacy is not one of degree but of orientation:

AI Pre-Literacy is not a synonym for any of these constructs. It names the pre-competency developmental territory that precedes them.

2.6 Scope Conditions

Typology of AI. “AI” is not a monolith. A conversational agent, a generative image model, a recommendation algorithm, a robotic toy, and a nursery monitoring system are developmentally distinct encounters. A child’s mental model of Alexa does not transfer to TikTok’s algorithm. AI Pre-Literacy addresses child-facing and child-adjacent AI-mediated phenomena: systems that shape, generate, curate, or simulate content, agency, or affect encountered by the child, whether directly or through the caregiver-mediated ecology. Systems that do not enter the child’s developmental ecology are outside the construct’s scope.

Caregiver mediation. In the 0–8 period, the caregiver is the primary mediator of AI exposure. The construct treats caregiver mediation as a contextual moderator, not a pillar. The child’s AI Pre-Literacy develops within, not apart from, the caregiver’s own AI stance, mediation style, and the parent–child AI interaction. This is a scope condition, not a fifth dimension of the child.

3. Intellectual Genealogy

AI Pre-Literacy is grounded in six research traditions, each of which contributes a distinct dimension to the construct.

The first is emergent literacy, which established that children are from birth “in the process of becoming literate” (Teale & Sulzby, 1986, p. xix), that literacy has developmental precursors (Whitehurst & Lonigan, 1998), and that deficit models of “pre-reading” were theoretically and empirically untenable (Clay, 1966; Teale & Sulzby, 1986).

The second is sociocultural and constructivist developmental theory, which holds that capability develops through graduated apprenticeship under adult scaffolding (Vygotsky, 1978; Bruner, 1990) and that children actively construct explanatory models of their world (Piaget, 1952).

The third is ecological systems theory, which situates development within nested, interacting systems in which homes, preschools, peer groups, media environments, languages, and socioeconomic conditions are constitutive rather than incidental (Bronfenbrenner, 1979; Bronfenbrenner & Morris, 2006).

The fourth is New Literacy Studies, which reframes literacy as a social practice rather than a neutral skill (Street, 1984) and recognizes that children bring funds of knowledge to all sense-making (Moll et al., 1992).

The fifth is critical AI studies, which situates AI within racial, economic, and infrastructural power (Benjamin, 2019; Noble, 2018; Crawford, 2021; Zuboff, 2019).

The sixth is posthumanist and relational accounts of childhood–technology relations, which caution against policing the human–AI boundary too rigidly and open inquiry into hybrid and distributed forms of agency (Turkle, 2011; Haraway, 2016; Braidotti, 2013).

Two further traditions inform the construct’s normative philosophy. Critical pedagogy (Freire, 1970) and constructionism (Papert, 1980) ground the claim that the aim of development is authorship, not consumption — the child as subject who acts on the world rather than as object acted upon.

4. Deliberate Elements of the Definition

Seven elements of the formal definition are deliberate and warrant explication.

First, pre as prime, not before. The prefix is deliberate. We do not use it in the deficit sense of “pre-reading” that Clay (1966) and Teale and Sulzby (1986) rightly overturned. We use it in the sense of prime: the first, foremost, formative years. AI Pre-Literacy names not a lack but a peak of developmental consequence — the period in which the orientations that will shape all later AI sense-making are laid down.

Second, developmental territory, not stage or competency. Following Teale and Sulzby (1986), we describe AI Pre-Literacy as a bounded region of inquiry rather than a universal sequence, a single latent capacity, or a necessary transition. This avoids the deficit logic of “pre-reading” that the emergent literacy literature has rightly criticized.

Third, seven developmental verbs. Encounter, notice, represent, interpret, discuss, orient toward, and respond map the full range from peripheral exposure to active sense-making and behavioral response, consistent with ecological models of proximal processes (Bronfenbrenner & Morris, 2006).

Fourth, foundations, not competencies in miniature. AI Literacy asks what a learner knows and can do (Long & Magerko, 2020). AI Pre-Literacy asks how a child is beginning to construct the orientations through which AI becomes intelligible. Foundations are what make competencies possible (Whitehurst & Lonigan, 1998).

Fifth, five interdependent foundations. Linguistic, cognitive, affective, empathic, and socio-ethical. These correspond to the five pillars elaborated below.

Sixth, within ecologies. Following Bronfenbrenner (1979), homes, preschools, peer groups, media environments, languages, and socioeconomic conditions are constitutive, not incidental.

Seventh, prototypical range 0–8. The upper bound is set primarily by the onset of AI literacy itself. AI literacy frameworks begin at age eight (Long & Magerko, 2020; Ng et al., 2021a, 2021b; Touretzky et al., 2019; UNESCO, 2021, 2023), and AI Pre-Literacy names the territory immediately preceding that threshold. Three converging markers support this boundary: second-order theory of mind consolidates at approximately seven to eight years (Wellman et al., 2001); UNESCO (2021) defines early childhood up to eight years as the learning-to-read period; and UNICEF (2021) and Common Sense Media (2023) issue AI guidance for the 0–8 range. The range is prototypical, not maturational: it marks where AI literacy takes over, not where development ends.

5. The Five Pillars of AI Pre-Literacy

The framework identifies five overlapping, non-sequential pillars. They are provisional dimensions for inquiry, not a validated measurement model. Each is grounded in peer-reviewed research on child development.

From Pillars to the Philosophy: How Mastery Emerges

The five pillars describe the dimensions of AI Pre-Literacy; the philosophy describes its aim. The connection between them is that each pillar contributes a distinct capacity for developmental sovereignty.

Together, the pillars describe what mastery rests on. None of them is mastery itself; mastery is what becomes possible when the five are cultivated together.

6. What AI Pre-Literacy Is Not

It is not AI literacy for younger children, not a readiness test, not a deterministic stage model (Bronfenbrenner & Morris, 2006), and not technocentric. It is not a replacement for emergent, digital, media, or critical literacy. It is not a competency framework. Following Teale and Sulzby (1986), there is no point at which it begins or ends in a maturational sense; it is bounded by encounter with AI-mediated phenomena.

7. Anticipated Objections and Clarifications

8. Frequently Asked Questions

9. Falsifiability and Research Agenda

The construct generates testable hypotheses. First, the five pillars will show discriminant validity: Pillar 3 (trust) will predict information-seeking, while Pillar 5 (empathic attunement) will predict peer-directed prosociality; they will not collapse into a single factor. Second, Pillar 2 mental models will show a qualitative shift at eight to nine years, corresponding to second-order theory of mind consolidation (Wellman et al., 2001). Third, Pillar 5 will predict lower over-disclosure to AI and higher peer-directed prosociality after AI interaction. This hypothesis is derived from the construct’s theoretical logic rather than from existing empirical evidence and awaits direct testing. Fourth, pillars will vary systematically by home mediation, language ecology, and socioeconomic context (Bronfenbrenner & Morris, 2006; Livingstone & Blum-Ross, 2020). Fifth, AI Pre-Literacy indicators will predict later AI literacy beyond emergent literacy and oral language (Whitehurst & Lonigan, 1998). If these hypotheses fail, the construct is unsupported as specified.

10. Limitations and Future Work

This is a conceptual introduction, not an empirical validation. Four limitations are acknowledged.

Attention is not yet a pillar. The construct does not include a dedicated attentional dimension, though attention is the substrate on which all AI Pre-Literacy develops and the dimension most conspicuously absent from existing AI literacy frameworks. Attention is reserved for a subsequent paper.

Play is not yet theorised as mechanism. Play is the primary developmental medium of the 0–8 period and appears in this framework as a method (play-based observation) rather than as the mechanism through which AI Pre-Literacy develops. This requires development.

Individual differences are not yet specified. The construct does not yet model variance by temperament, prior exposure, parental mediation style, or child resistance and refusal. A variance theory is needed.

Pillar 5 lacks direct empirical support. The empathic attunement pillar is the construct’s most distinctive claim and its most theoretically driven. Direct empirical research on displacement effects and empathic differentiation in child–AI interaction is limited; the pillar awaits operationalisation and testing.

Next steps: (1) ethnographic documentation in cross-cultural sites; (2) development of a play-based observational protocol; (3) longitudinal study testing whether Pillar 5 predicts lower over-disclosure and higher peer prosociality; (4) participatory design with children, parents, and teachers to co-define appropriate trust; (5) a follow-up paper developing attention as a sixth pillar.

11. Glossary

AI Pre-Literacy

The developmental territory from birth to approximately eight years in which children form early linguistic, cognitive, affective, empathic, and socio-ethical foundations for AI sense-making.

AI-mediated phenomena

Experiences in which AI systems shape, generate, curate, or simulate content, agency, or affect encountered by the child.

Child-oriented approach

An approach that begins from the developing person and asks how the conditions of development have changed. Characteristic of AI Pre-Literacy.

Developmental sovereignty

The capacity to author with AI rather than being authored by it: to interrogate, direct, refuse, and hold AI systems to account. The normative aim of AI Pre-Literacy.

Empathic differentiation

The capacity to hold connection to AI together with recognition that simulated affect is not human feeling.

Epistemic trust calibration

The developmental process of learning when to believe, question, or distrust AI systems.

Foundations

Pre-competency orientations that make later competencies possible; not competencies in miniature.

Mastery orientation

Orientation toward AI as a domain one can interrogate and direct, rather than as a service one consumes. Contrasted with consumer orientation.

Object-oriented approach

An approach that begins from the technology and asks what learners should know and be able to do. Characteristic of AI literacy frameworks.

Pillars

Five provisional dimensions of AI Pre-Literacy: vocabulary, sense-making, trust, socio-ethical awareness, empathic attunement.

Pre- (as used in AI Pre-Literacy)

Taken in the sense of prime: the first, foremost, formative years. Not a deficit prefix indicating absence or beforeness.

Structural noticing

Adult-scaffolded awareness that AI systems are embedded in commercial, racial, and infrastructural power.

Territory

A bounded region of developmental experience, following Teale and Sulzby (1986); not a stage or latent trait.

12. How to Cite

13. References

Changelog

Version 2.0 (September 20, 2026). Expanded formal statement of the construct. Adds empathic attunement as a fifth pillar, specifies a prototypical range of birth to eight years, and makes explicit the reading of pre as prime. First public use of the term: August 19, 2026 (Pasha & Pasha, 2026; https://doi.org/10.5281/zenodo.22023810). Version history, changelog, and pending work are maintained on the Zenodo record for this deposit.