A researcher opens six databases. She types “young children and AI” into the first, “toddlers and smart speakers” into the second, “preschoolers and social robots” into the third. By the fourth search she is no longer sure she is looking at the same field — or the same phenomenon.
Four hours later, she has a folder of PDFs and no map.
This is the problem the AI Pre-Literacy Research Observatory was built to solve. It is a free, open, searchable evidence map of research on young children’s encounters with artificial intelligence and intelligent agents. Not a bibliography — an instrument.
Explore the Observatory: aipreliteracy.org/research-observatory
The Problem
The field of early childhood and AI has a visibility problem, not an activity problem.
Research is being produced across disciplines that rarely cite one another — early childhood education, human–robot interaction, developmental psychology, media studies, computer science, ethics, and policy. Each has its own vocabulary for describing the same phenomenon. A “conversational agent” in one paper is a “voice assistant” in another and a “social robot” in a third.
For a researcher entering this space, the consequences are costly:
- Terminology fragmentation. A keyword search returns a fraction of what exists, because the same construct carries many names.
- Disciplinary silos. Education researchers and HCI researchers study adjacent questions without reading each other.
- Blurred evidence boundaries. Primary findings sit alongside commentary and opinion, looking equivalent in a flat list.
- Unresolved questions presented as settled. Genuine disagreement gets smoothed away in summary.
- No stable search surface. Because the domain has no name, it has no conference track, no funding category, no standard search string.
Studies are being produced. Knowledge is not accumulating in a way anyone can see.
Our Effort: The Research Observatory
The solution is not another literature review — reviews are snapshots, and they date. The solution is evidence infrastructure: a structured, maintained map that makes the shape of a research landscape visible and inspectable.
What the Observatory contains:
- 62 discovery records spanning the emerging literature on young children and AI
- Three analytical layers separating records by their function in the evidence base
- A structured, multi-concept discovery search — essential in a domain where one phenomenon carries many names
- Row-level verification, so every record can be inspected individually rather than accepted as an opaque aggregate
- Explicit analytical boundaries keeping primary evidence distinct from contextual, emerging, and unresolved records
Two stated methodological positions. The Observatory is explicit about two decisions that shape its contents:
- It does not treat all social robots as AI. A robot responding to a predetermined script is not the same object of study as an AI system. Conflating them inflates the evidence base and misrepresents what children actually encounter.
- It does not use school grade as a proxy for chronological age. Grade levels vary by country and enrolment timing. Age is the developmentally meaningful variable, and it is recorded as such.
These positions are stated openly, not buried in a methods appendix. A map with visible rules can be verified, cited, and disagreed with. A map with hidden rules can only be trusted or ignored.
Preserved original terminology. The Observatory records what studies actually called things. It does not silently rename a construct to fit a framework that did not exist when the study was designed.
Why “AI Pre-Literacy” — A New Domain
Most discussion of children and AI jumps directly to AI literacy: what children should eventually understand about algorithms, data, and automated systems.
But there is a period before that. Before a child can read a prompt, before they can question a recommendation, before they have any concept of what a machine is — they are already interacting with AI. They are forming expectations about whether objects respond, whether voices are persons, whether a screen that answers is a thing that knows.
That period is AI pre-literacy: the formative stage of children’s encounters with AI and intelligent agents, prior to formal AI literacy.
We are promoting it as a research domain for four reasons:
1. The exposure is real and it is early. These encounters are happening whether or not there is a name for them. A domain without a name is a domain without a literature, a conference track, a funding category, or a curriculum.
2. It is conceptually distinct from adjacent fields. It is not “AI literacy, but younger.” It is not “screen time.” It is not child–robot interaction alone. It sits at the intersection and requires its own framing.
3. It is a policy-relevant window. Decisions about early years technology are being made now, largely without an evidence base. A named domain creates a place to put evidence when it arrives.
4. Naming a field is how fields become visible. Researchers search for terms. Reviewers write about terms. Funders fund terms. If the domain has no name, it has no surface area.
Critically, the Observatory introduces AI pre-literacy only at the synthesis level, as a provisional organizing domain — not as a claim about what the underlying studies were studying. The literature comes first. The label serves the literature; it does not replace it.
Who Benefits
Researchers. If you are planning a study, writing a review, positioning a grant, or looking for a gap, the Observatory shortens the distance between a question and a defensible answer. It is also designed to be disagreed with — the analytical boundaries are visible so they can be contested.
Doctoral and early-career researchers. Entering a fragmented, cross-disciplinary domain is hardest at the beginning. A structured map with explicit classification decisions is a faster, fairer on-ramp than a keyword search across six databases.
Educators and early years practitioners. If you are making decisions about technology in an early years setting, you deserve to know what the evidence supports — and where it runs out.
Policymakers and regulators. When the question is “what do we know about very young children and AI systems?”, the honest answer should include a clear boundary between established findings and unresolved questions. That boundary is built into the structure.
Developers and designers. If you are building products used by children who cannot yet read, the research base is your constraint and your guide.
Funders and research strategy leads. A mapped landscape is a map of under-invested areas.
New Dimensions and Opportunities for Researchers
A named domain does more than organize existing work. It opens new territory.
For young researchers, AI pre-literacy offers rare advantages. It is a field with low entry barriers — the literature is still small enough to read comprehensively, and foundational questions remain genuinely open. Early-career scholars can make visible, citable contributions quickly: a scoping review, a terminology mapping, a small qualitative study. There is also room to shape the vocabulary itself. In an established field, a doctoral student inherits a framework. In a forming field, they can help build one.
For senior researchers, the domain offers something different: a place to bring mature methodological expertise into unmapped territory. Scholars with established work in developmental psychology, HCI, ethics, or education policy can extend their frameworks into questions that have not yet been asked well. Cross-disciplinary collaboration is not optional here — it is the native condition of the field. Senior researchers are also positioned to do what the domain most needs: build the rigorous longitudinal work that early-stage scholars cannot yet lead.
For both, the opportunities are structural:
- Genuine gaps. The evidence base is thin enough that meaningful gaps are visible and addressable.
- Cross-disciplinary reach. Work in this domain is citable across education, psychology, computer science, and policy.
- Policy relevance. Early years AI decisions are being made now. Research that arrives in this window has unusual influence.
- International comparative work. Frameworks from the OECD, UNESCO, and UNICEF all point toward early foundational awareness, yet cross-national evidence remains scarce.
The field is young enough that a single well-designed study can define a sub-area. That window does not stay open long.
A Positive Conclusion
The AI Pre-Literacy Research Observatory is not a finished statement about a settled field. It is a working instrument for a field still forming.
That is the point. A domain in its early years needs infrastructure more than conclusions — a way to see what exists, what it means, and where the gaps are. The Observatory is free, open, and requires no login or institutional affiliation. It was built for the research community, and it will improve fastest if that community uses it and pushes back on it.
The evidence base for young children and AI is growing. Now it has a map — and the territory beyond it is open.
About the Authors
Dr. Shaheen Pasha — Retired Professor of Special Education. Ph.D., University of Southampton, UK. Former Chairperson, Department of Special Education, University of Education, Lahore. 35+ research papers, 2 books.
Dr. M. A. Pasha — Retired Professor of Computer Science. Ph.D. in Artificial Intelligence, University of Southampton, UK (1996). Research: AI, Human-Computer Interaction, Computational Thinking. 30+ research papers, 2 books.
Their mission: 70 years of combined academic wisdom, united to foster AI pre-literacy in young children.