1. Introduction and problem

In the 3–6 age band—kindergarten, preschool, CENDI, early childhood school—a package of five artefacts has been installed that purport to count as support for play with loose parts / loose parts play / open-ended materials: a chatbot that “suggests constructions” or “loose parts ideas” to the child; a GenAI generator of “activities” or “provocations” by prompt that fixes the outcome; a computer-vision system that scores “creativity” or labels assembly photos; an AR layer that “guides” how to arrange materials; and a dashboard of “exploration metrics” or “open-endedness score.” All five allow centres to display that they “already do loose parts with AI.” The leap—from suggestion, provocation, score, AR guidance or metric to affirming support—is not authorized by AI-in-ECE mappings nor by loose-parts pedagogy when there are open-ended physical materials, child agency, prolonged time, peer negotiation, adult co-presence without a predetermined outcome and divergent exploration.

This article’s thesis is restrictive. A chatbot that “suggests constructions” or “loose parts ideas” to the child, a GenAI generator of “loose parts activities” or “provocations” by prompt, a computer-vision system that scores “creativity” or labels assembly photos, an AR layer that “guides” how to arrange materials, or a dashboard of “exploration metrics” / “open-endedness score” do not constitute support for loose parts play in the early years. In early childhood, loose parts play develops with open-ended physical materials, child agency, prolonged time, peer negotiation and adult co-presence without a predetermined outcome; not a creativity score, an AI suggester, a prompt-generated provocation that fixes the outcome, or an AR layer that directs arrangement. Cankaya, Rohatyn-Martin, Leach, Taylor and Bulut (2023) review the literature on loose parts play and cognitive development in preschool and document an empirical gap, especially indoors. Cankaya, Martin and Haugen (2025) apply PRISMA to 25 studies through December 2024 (ages 0–6): only one study explicitly uses the term “loose parts”; they report associations with problem-solving, creativity and skills, with methodological gaps. Cankaya, Rohatyn-Martin, Buro, Bulut and Taylor (2025)—N = 60; Mage = 58.6 months (SD = 10.9); within-subjects—find more spontaneous STEM behaviours with loose parts than with limited-function toys: an empirical finding of open affordances, not authorization to turn this article into one on curricular STEM/robotics. Aşkar and Durmuşoğlu (2023) contribute a qualitative case on the meaning of play with loose parts in preschool. Zeng and Ng (2025) document human questioning strategies during loose parts play in kindergarten. That authorizes asking what was measured: the bot’s suggestion, a prompt provocation, a CV label, AR guidance or an open-endedness score—not the practice when a child reconfigures an open material, negotiates with a peer, sustains the episode in time and receives adult co-presence without a predetermined outcome.

The problem is aggravated by five category confusions. First: loose parts / open-ended materials is not generic “free play”—although free play may host it, the axis is the practice of open affordances with physical loose parts. Second: it is not block play / spatial reasoning as axis—loose material may include reconfigurable pieces, but it is not reducible to measurable spatial assembly. Third: it is not outdoor play / nature play—Hu (2025) contributes nature-based loose parts only as an affordance contrast, not as the already-published outdoor axis. Fourth: it is not curricular artistic creativity nor a creativity score. Fifth: it is not STEM/robotics as axis—Cankaya et al. (2025) are read only as evidence of spontaneous affordances. This work does not recycle outdoor play, free play, symbolic play, block play/spatial, STEM/robotics, artistic creativity, motor development, environmental education, inquiry, literacy, orality, SEL, executive functions, UDL, documentation, formative assessment, participation, planning, joint attention, adult–child interactions as generic axis, teacher education, music or numeracy. The question is what counts as support for loose parts play when a centre “does AI and loose parts.”

There is, moreover, an economy of coordination: suggester chatbot, GenAI provocation, creativity score, AR layer and open-endedness dashboard fit on a slide; the prolonged episode with open materials, child agency, peers and adult co-presence without a fixed outcome does not. NAEYC (2022) and OECD (2021, 2023) require meaningful interactions and subordinated digitization. Inference: support for loose parts play is not fulfilled by generating provocations or scoring exploration. Contributions are three: separating craft from the artefact package; examining three case families—including Zeng and Ng (2025) as human scaffolding through questioning and Lyu and McNair (2026) as peripheral contrast (AI→adult)—; and offering four tests to affirm support for loose parts play, not merely chatbot, GenAI, CV, AR or dashboard.

2. State of the art: from situated loose-parts practice to the artefact on display

It is useful to separate four strata that the market of “AI for loose parts in ECE” tends to mix. The first is the construct of loose parts / loose parts play / open-ended materials at ages 3–6 as practice of open affordances with reconfigurable physical materials, child agency, prolonged time, peer negotiation and divergent exploration without a predetermined outcome (Cankaya et al., 2023; Cankaya, Martin & Haugen, 2025; Aşkar & Durmuşoğlu, 2023; Zeng & Ng, 2025; Hu, 2025). The second is the pedagogical craft that cultivates it—offering open materials, protecting time, human questioning that opens without fixing a result, adult co-presence, not replacing the child’s reconfiguration with a digital suggestion—(Zeng & Ng, 2025; Aşkar & Durmuşoğlu, 2023; Lyu & McNair, 2026; NAEYC, 2022; OECD, 2021, 2023). The third is evidence on AI affordances in ECE, child-centred GenAI and Froebel+AI dual-track—without equating them to situated loose parts play with physical materials—(Chen, 2024; Su & Yang, 2022; Su & Zhong, 2022; Ljungcrantz, 2026; Nikolopoulou, 2025; Lyu & McNair, 2026). The fourth is the rights, systems and developmentally appropriate practice framework that treats the 3–6-year-old as a subject of divergent exploration with open materials, not as a signal vector for an open-endedness-score pipeline nor as a passive recipient of chatbot suggestions or a GenAI provocation (UNESCO, 2021; Miao & Holmes, 2023; U.S. Department of Education, 2023).

In construct and craft, Cankaya et al. (2023) synthesize LPP–cognition in preschool and mark an indoor empirical gap: review finding. Cankaya, Martin and Haugen (2025)—PRISMA; 25 studies; 0–6—confirm sparse explicit use of the term “loose parts” and associations with problem-solving, creativity and skills alongside gaps: systematic-review finding. Cankaya et al. (2025)—N = 60; Mage = 58.6 m (SD = 10.9)—document more spontaneous STEM behaviours with loose parts than with limited toys: affordance finding, not a STEM axis. Aşkar and Durmuşoğlu (2023) contribute qualitative meaning. Zeng and Ng (2025) show human questioning during LPP without a fixed outcome. Hu (2025) infuses nature-based loose parts via action research: contrast with already-published outdoor. Inference: a chatbot does not reconfigure the piece; a child with a peer and an adult who asks without fixing the result does.

In the artefact stratum, Chen (2024), Su and Yang (2022), Su and Zhong (2022) and Ljungcrantz (2026) map AI affordances, curricula and state of the art in ECE without equating them to situated loose parts play. Nikolopoulou (2025) balances promises and challenges of child-centred GenAI under teacher mediation: caution framework. Lyu and McNair (2026) reconceptualize Froebelian pedagogy with AI in a Chinese context (N = 50 practitioners): dual-track integration; AI is positioned peripherally to the adult; concerns emerge about displacement of sensory experiences. In the systems stratum, UNESCO (2021) requires human oversight. Miao and Holmes (2023) set pedagogical validation and age thresholds for generative AI—direct framework for the prompt-based provocations generator. U.S. Department of Education (2023) requires that AI support, not replace, professional judgement. OECD (2021, 2023) anchor meaningful interactions and subordinated digitization; NAEYC (2022) anchors developmentally appropriate practice.

3. Review method

A critical narrative review was conducted, not a primary meta-analysis. The purpose was not to estimate a homogeneous effect size of the five artefacts, but to articulate a pedagogical category argument with verified sources. Inclusion criteria: (a) 2021–2026, with transfer marked when the sample does not equal ages 3–6; (b) loose parts play, open-ended materials, loose parts, open affordances, or AI in ECE relevant to artefact/craft contrast; (c) kindergarten, preschool, CENDI or ages 3–6; (d) peer-reviewed journal, DOI or NAEYC/UNESCO/OECD report; (e) verifiable DOI or publisher page. Axes already used in this series were excluded as central object—outdoor play, free play, symbolic play, block play/spatial, STEM/robotics, artistic creativity, motor development, environmental education, inquiry, literacy, orality, SEL, executive functions, UDL, documentation, formative assessment, participation, planning, joint attention, adult–child interactions as generic axis, teacher education, music and numeracy. STEM behaviours in Cankaya et al. (2025) appear only as spontaneous affordances; Hu (2025) only as contrast with outdoor.

The search was executed on 3 September 2026 (slot 01:02 America/Mexico_City) on DOI pages, Crossref, Springer, Elsevier, Taylor & Francis, Nature Portfolio, MDPI, JAIR, OECD iLibrary, UNESDOC, NAEYC and publisher sites. Each source was verified. Empirical finding, framework and pedagogical inference were distinguished. No N, d, r, AUC or DOI was invented: when an artefact lacks a verified study with scoring metrics at ages 3–6, it is discussed as a category ceiling supported by AI-in-ECE mappings (Chen, 2024; Su & Yang, 2022; Ljungcrantz, 2026) and by the Froebel+AI dual-track (Lyu & McNair, 2026). The eighteen sources of the verified corpus were used.

4. Case 1. Construction/loose-parts-idea suggester chatbot, prompt-based GenAI provocations, CV creativity score, arrangement-guiding AR or open-endedness dashboard do not constitute support for loose parts play

Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) saturate the artefact-ceiling portrait when AI in ECE is presented as if it were support for loose parts play. Chen (2024) maps global AI affordances in early childhood education: scoping finding on emerging uses—tutoring, analytics, content generation—not a finding that a chatbot suggesting constructions cultivates child agency with open-ended physical materials. Su and Yang (2022) review the AI-in-ECE field: synthesis finding on trends, not on peer negotiation with reconfigurable pieces or prolonged divergent exploration. Ljungcrantz (2026) reviews AI–ECE interaction 2020–2024: state-of-the-art finding, not situated loose parts play. Pedagogical inference, marked as such: the gesture “the child received chatbot ideas = there was loose parts play” is a generative-tutoring ceiling. Cankaya et al. (2023), Cankaya, Martin and Haugen (2025) and Aşkar and Durmuşoğlu (2023) require open materials, agency and situated meaning; a useful dialogue with a bot can coexist with absence of physical materials, peers, protected time and an open outcome.

Nikolopoulou (2025) and Miao and Holmes (2023) name the risk of the generator of “loose parts activities” or “provocations” from a prompt that fixes the outcome. Nikolopoulou (2025) balances promises and challenges of child-centred GenAI under teacher mediation: caution framework. Miao and Holmes (2023) require pedagogical validation and age thresholds for generative AI. Status: normative and review framework, not a trial of prompt provocations versus a loose-parts episode with human questioning (Zeng & Ng, 2025). Inference: producing a provocation by prompt may be subordinated teacher preparation; support for loose parts play begins when there are open materials on the table or floor, child agency to reconfigure, protected time, negotiating peers and an adult who asks without fixing the result. Su and Zhong (2022) propose AI curriculum design in ECE as future direction—AI literacy curriculum, not situated loose parts play. Boundary inference: an AI curriculum does not sign divergent exploration with loose parts.

The computer-vision system that scores “creativity” or labels assembly photos, the AR layer that “guides” how to arrange materials and the dashboard of “exploration metrics” / “open-endedness score” lack, in the verified corpus, trials with reported N, d, r or AUC for loose-parts creativity scoring, AR arrangement guidance or open-endedness score in preschool 3–6; figures are not invented. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map analytics, tutoring, agents and generation as AI-in-ECE trend, without equivalence to mediated loose parts play. Cankaya et al. (2025) measure spontaneous STEM behaviours with loose parts versus limited toys (N = 60): affordance finding, not algorithmic creativity score. Lyu and McNair (2026)—N = 50 Froebelian practitioners—document concerns about displacement of sensory experiences when AI approaches the child: empirical contrast of legitimate peripheral use only when AI supports the adult. Restrictive inference: GenAI supporting the teacher ≠ suggester chatbot to the child; prompt provocation that fixes the outcome ≠ human questioning that opens (Zeng & Ng, 2025); teacher empowerment ≠ open-endedness score; AR layer that directs arrangement ≠ free reconfiguration of the material. UNESCO (2021) and U.S. Department of Education (2023) require human oversight. The five artefacts share one substitution grammar: the bot’s suggestion speaks for the child’s agency; the GenAI provocation fixes the outcome; the CV score replaces pedagogical observation of the episode; AR directs arrangement; the dashboard replaces prolonged time with a metric.

It is useful to specify the ceiling without inventing effects. A chatbot can list “ideas”; Aşkar and Durmuşoğlu (2023) and Zeng and Ng (2025) document human meaning and questioning, not turns with a model. A GenAI can print a provocation; Miao and Holmes (2023) require pedagogical validation. A CV system can label “high creativity”; Cankaya, Martin and Haugen (2025) review human associations, not AUC. An AR layer can guide arrangement; Cankaya et al. (2025) measure spontaneous behaviours, not digital guidance. A dashboard can raise an “open-endedness score”; NAEYC (2022) and OECD (2021) anchor practice and interactions. Inference: “high creativity score” or “provocation per the model” is measurement or generation, not signed support.

5. Case 2. What the kindergarten does when there is support for loose parts play: open materials, agency, time, peers, co-presence—with Zeng and Ng (questioning) and Lyu and McNair (dual-track) as contrast

Cankaya, Martin and Haugen (2025) saturate the empirical review floor: 25 studies through December 2024 (ages 0–6; PRISMA) associate indoor loose parts play with problem-solving, creativity and skills, with the caveat that only one study explicitly uses “loose parts” and methodological gaps persist. Status: systematic-review finding. Marked transfer: the 0–6 range includes toddlerhood; it is read with caution toward the kindergarten 3–6 band. Cankaya et al. (2023) precede the indoor LPP–cognition gap in preschool: review finding. Pedagogical inference, marked as such: this is the kind of object an early childhood setting can call support for loose parts play when it protects open-ended physical materials, child agency and time—not when it scores open-endedness on a dashboard. Loose parts play is not a creativity score: it is growing command of reconfiguring open materials, sustaining the episode, negotiating with peers, exploring divergently and receiving adult co-presence without a predetermined outcome. A chatbot that suggests “make a bridge” without shared materials can celebrate a labelled dialogue and, at once, empty agency. A CV system can label “creative assembly” in a photo and not have captured whether the adult fixed the outcome, whether there were peers, or whether the material was truly open.

Aşkar and Durmuşoğlu (2023) document, in a preschool qualitative case, the meaning of play with loose parts: craft is read from situated experience, not from a score. Zeng and Ng (2025) contribute the decisive human scaffolding: questioning strategies during loose parts play in kindergarten that promote science learning without converting free play into a lesson with a closed outcome. Status: empirical finding of mediation by human questions. Inference: Zeng and Ng’s questioning is the direct contrast with a GenAI prompt provocation that fixes the result—the adult opens; the prompt closes. Cankaya et al. (2025)—N = 60; Mage = 58.6 m (SD = 10.9); within-subjects—show more spontaneous STEM behaviours with loose parts than with limited toys: evidence of open affordances; it is not converted here into a curricular STEM/robotics axis. Hu (2025) shows action research that infuses nature-based loose parts into a Kindergarten Program: natural loose material as affordance, with explicit contrast versus outdoor play already published in the series. NAEYC (2022) and OECD (2021, 2023) anchor meaningful interactions and subordinated digitization.

The contrast of peripheral good use is Lyu and McNair’s (2026) Froebel+AI dual-track. They report N = 50 Froebelian practitioners in a Chinese context: dual-track integration; AI is conceived as peripheral to the adult/teacher; concerns emerge about displacement of sensory experiences if technology approaches the child as substitute. Status: empirical finding of pedagogical reconceptualization and practitioner concern. Inference: subordinated digital—adult receiving AI support to plan or reflect without replacing physical materials or the child’s agency—can be coherent with loose parts play; it does not authorize declaring that a suggester chatbot, a GenAI provocation that fixes the outcome, a CV creativity score, an AR layer that guides arrangement or an open-endedness score constitute the craft. Mediation—NAEYC (2022), OECD (2021); Zeng and Ng (2025)—is adult co-presence that organizes time, offers open materials, asks without fixing a result, listens to peer negotiation and expands without imposing a digital script. Inference: pedagogical observation of loose parts play is professional reading of the episode with open materials; the dashboard counts a metric. A dashboard does not reconfigure the piece; a child and an adult who asks without fixing the outcome do.

6. Case 3. Loose parts ≠ generic free play, spatial block play, outdoor-nature play, curricular artistic creativity or STEM-robotics as axis

The first category boundary is generic free play. Free play may host loose parts; it does not, by itself, constitute the practice of open affordances with reconfigurable physical materials as axis. Inference: this article forbids the equivalence “we have free play = there is support for loose parts play.” A chatbot that “suggests ideas” is not free either: it is generative tutoring that narrows agency. The second boundary is already-published block play / spatial reasoning: 2D/3D assembly with unit blocks and spatial language is a distinct construct; loose material may include pieces, but the axis here is openness of affordances, not a spatial score or block talk as object. The third is outdoor play / nature play: Hu (2025) contributes nature-based loose parts as infusion of natural loose materials—affordance contrast—not identity with green time, forest school or outdoor minutes already treated. Inference: “we have a yard = there is loose parts” does not sign open materials, agency and reconfiguration time. The fourth is curricular artistic creativity or the creativity score: Cankaya, Martin and Haugen (2025) report associations with creativity as a possible lateral outcome in human reviews; they do not authorize CV that scores creativity of assembly photos. The fifth is STEM/robotics as axis: Cankaya et al. (2025) document spontaneous STEM behaviours with loose parts—affordance finding—; converting them into a STEM or robotics curriculum inverts this article’s argument. UNESCO (2021), Miao and Holmes (2023) and U.S. Department of Education (2023) require human oversight. Inference: the only AI use coherent with ages 3–6 remains on the adult’s side—Lyu and McNair’s (2026) Froebel+AI dual-track—subject to pedagogical validation and contrasted with human questioning (Zeng & Ng, 2025). It does not enter as autonomous scorer, suggester chatbot, GenAI provocation that fixes the outcome, AR that directs arrangement or open-endedness dashboard.

7. Inferential framework: four tests to affirm support for loose parts play, not an artefact

The framework that follows is this article’s pedagogical inference, anchored in the cases and verified instruments. It is not a new international standard. It distinguishes four tests. If a kindergarten, preschool, CENDI or early childhood school fails them, it cannot declare that the five artefacts constitute support for loose parts play.

7.1. Test of situated practice with open-ended physical materials, child agency and divergent exploration—not of the suggester chatbot or the AR layer. Cankaya et al. (2023), Cankaya, Martin and Haugen (2025), Aşkar and Durmuşoğlu (2023) and Cankaya et al. (2025) define the craft as observable loose parts and affordances. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map chatbots and content generation as AI affordances, not as loose parts play. Inference: evidence of support for loose parts play is verified in whether the child reconfigured open-ended physical materials with agency. If the centre’s “evidence” is a chatbot dialogue or AR arrangement guidance, the centre has done tutoring or digital direction of the material, not support for loose parts play.

7.2. Test of prolonged time, peer negotiation and human questioning without a fixed outcome—not of the GenAI prompt provocation. Zeng and Ng (2025), Aşkar and Durmuşoğlu (2023) and NAEYC (2022) situate questioning, situated meaning and developmentally appropriate practice. Miao and Holmes (2023) require pedagogical validation of generative AI. Inference: producing a prompt provocation that fixes the result does not demonstrate protected time, peers or human questions that open. A script may exist; it does not sign the loose-parts episode.

7.3. Test of human mediation—including the peripheral Froebel+AI dual-track contrast—not of the open-endedness dashboard or the CV creativity score. OECD (2021, 2023) require meaningful interactions and subordinated digitization. Lyu and McNair (2026) anchor AI peripheral to the adult with concern about sensory displacement. Zeng and Ng (2025) anchor human questioning. Inference: a “high open-endedness score” or “creativity score” may raise a threshold without raising co-presence quality. Pedagogical observation reads the episode with open materials; the dashboard counts a metric. GenAI guiding the adult is legitimate peripheral use; a chatbot suggesting constructions to the child or AR directing arrangement is not.

7.4. Test of category distinction and professional judgement—not of the product catalogue. Loose parts ≠ generic free play, spatial block play, outdoor-nature play, curricular artistic creativity or STEM-robotics as axis. Cankaya et al. (2025) prevent reducing spontaneous affordances to a STEM curriculum. Hu (2025) prevents confusing nature-based LPP with outdoor play as axis. Cankaya, Martin and Haugen (2025) prevent confusing creativity associations in human reviews with a CV score. UNESCO (2021), Miao and Holmes (2023), U.S. Department of Education (2023), NAEYC (2022) and OECD (2021, 2023) require human oversight, pedagogical validation and not replacing professional judgement. Inference: a centre cannot treat the infant as an open-endedness vector nor as exclusive interlocutor of an AI suggester. Support for loose parts play is not fulfilled by scoring “exploration” better. It is fulfilled by practising open-ended physical materials, child agency, prolonged time, peer negotiation and adult co-presence without a predetermined outcome.

The framework admits digital subordinated to teacher preparation and validated human mediation (Lyu & McNair, 2026; Chen, 2024; Su & Yang, 2022; Ljungcrantz, 2026). It rejects declaring support via the five artefacts (Miao & Holmes, 2023; Nikolopoulou, 2025; UNESCO, 2021). The four tests are read together.

8. Discussion

Three tensions organize the discussion. The first is between displaying suggester chatbot, GenAI provocations, CV score, AR or dashboard and exercising support for loose parts play. It is a finding that LPP–cognition reviews, the PRISMA of 25 studies, the within-subjects N = 60, the qualitative case, human questioning, nature-based LPP and the Froebel+AI dual-track (N = 50) sustain the craft or its boundary (Cankaya et al., 2023; Cankaya, Martin & Haugen, 2025; Cankaya et al., 2025; Aşkar & Durmuşoğlu, 2023; Zeng & Ng, 2025; Hu, 2025; Lyu & McNair, 2026). It is a framework that AI in ECE grows without equivalence to situated loose parts play (Chen, 2024; Su & Yang, 2022; Ljungcrantz, 2026). It is not a finding that the five artefacts produce the craft: they suggest, provoke with a fixed outcome, score, guide or measure what engineering delivers.

The second is between automated assessment of “creativity” or “open-endedness” and everyday pedagogy of open materials. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) objectivize analytics as trend; they do not report verified trials of creativity or open-endedness scoring in preschool 3–6—metrics are not invented. Inference: insisting that the kindergarten “already supports loose parts play” because a model scores creativity or a chatbot suggests ideas is inverted pedagogy: the algorithmic proxy is made to stand for open materials, agency, time and peers.

The third is between GenAI provocation / AR guidance and integral craft. Miao and Holmes (2023) and Nikolopoulou (2025) set GenAI limits; Lyu and McNair (2026) warn of sensory displacement; Zeng and Ng (2025) fix human questioning that opens. Inference: selling a prompt provocation or AR arrangement guidance as “loose parts with AI” confuses textual product or digital direction with situated practice. Cankaya et al. (2025) confirm spontaneous affordances without authorizing a STEM axis; UNESCO (2021), Miao and Holmes (2023) and U.S. Department of Education (2023) subordinate AI to professional judgement.

The four tests in section 7 read these tensions with operational criteria. The empirical craft contrast—LPP reviews; N = 60; qualitative case; questioning; nature-based LPP; dual-track N = 50—defines the floor the five artefacts do not reach alone. Inference: support erodes not by using AI, but by treating chatbot, GenAI, CV, AR or dashboard as if they were the practice. Subordinated digital is legitimate—AI peripheral to the adult (Lyu & McNair, 2026)—; inverting the sequence is not. It is verified with open materials, agency, time, peers and co-presence without a fixed outcome (NAEYC, 2022; OECD, 2021, 2023; Zeng & Ng, 2025).

9. Limits

This review is narrative. It does not apply its own PRISMA nor estimate primary combined effects. Cankaya, Martin and Haugen (2025) do apply PRISMA to 25 studies (0–6; partial transfer toward 3–6) and document gaps. Cankaya et al. (2023) review LPP–cognition with an indoor gap. Cankaya et al. (2025) measure N = 60: STEM behaviours as affordances, not as this article’s STEM axis. Aşkar and Durmuşoğlu (2023) contribute a qualitative case; Zeng and Ng (2025), human questioning; Hu (2025), nature-based LPP as outdoor contrast; Lyu and McNair (2026), dual-track N = 50. Chen (2024), Su and Yang (2022), Su and Zhong (2022), Ljungcrantz (2026) and Nikolopoulou (2025) map AI in ECE, not loose-parts creativity scoring. No verified trials of CV creativity score, suggester chatbot, guiding AR or open-endedness dashboard in CENDI 3–6 were located; they are discussed as category ceiling. NAEYC, UNESCO and OECD are framework sources. Section 7 inferences are category hypotheses, not implementation evidence.

10. Conclusions

A chatbot that “suggests constructions” or “loose parts ideas” to the child, a GenAI generator of “loose parts activities” or “provocations” by prompt, a computer-vision system that scores “creativity” or labels assembly photos, an AR layer that “guides” how to arrange materials, or a dashboard of “exploration metrics” / “open-endedness score” do not constitute support for loose parts play in an early childhood centre. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) confirm AI affordances without equivalence to situated loose parts play. Nikolopoulou (2025) and Miao and Holmes (2023) set GenAI limits. Lyu and McNair (2026) require reading the Froebel+AI dual-track as peripheral contrast—AI→adult ≠ suggester chatbot, GenAI provocation, CV score, AR or dashboard. Zeng and Ng (2025) fix human questioning as scaffolding that opens without fixing an outcome. When there is support, there is situated practice: LPP–cognition reviews and PRISMA of 25 studies (Cankaya et al., 2023; Cankaya, Martin & Haugen, 2025); spontaneous affordances N = 60 (Cankaya et al., 2025; STEM contrast); qualitative meaning (Aşkar & Durmuşoğlu, 2023); nature-based LPP as outdoor contrast (Hu, 2025); DAP and interactions (NAEYC, 2022; OECD, 2021, 2023). Loose parts play is distinguished from generic free play, spatial block play, outdoor-nature play, curricular artistic creativity and STEM-robotics as axis. Guidance documents require human oversight (UNESCO, 2021; Miao & Holmes, 2023; U.S. Department of Education, 2023; Su & Zhong, 2022).

Where sources do not measure a kindergarten, this article does not affirm it. Where they measure AI mappings, GenAI or practitioner dual-track, it does not translate them into pedagogical support for loose parts play. Accompanying three- to six-year-olds in loose parts play is to practise open-ended physical materials, child agency, prolonged time, peer negotiation, divergent exploration and adult co-presence without a predetermined outcome. The rest is suggester chatbot, GenAI prompt provocation, creativity scoring, AR arrangement guidance and open-endedness dashboard. It is not support for loose parts play in early childhood education, and it must not be presented as what it is not.

Laboratorio Editorial de NEXTECH.IA / Ingeniero Mitre.

References

  1. Aşkar, N., y Durmuşoğlu, M. C. (2023). Meaning of play with loose parts materials in preschool education: A case study. Journal of Qualitative Research in Education, (33), 71–111. https://doi.org/10.14689/enad.33.1645
  2. Cankaya, O., Martin, M., y Haugen, D. (2025). The relationship between children’s indoor loose parts play and cognitive development: A systematic review. Journal of Intelligence, 13(5), 52. https://doi.org/10.3390/jintelligence13050052
  3. Cankaya, O., Rohatyn-Martin, N., Buro, K., Bulut, O., y Taylor, K. (2025). Loose parts play encourages spontaneous science, technology, engineering, and mathematics (STEM) behaviours. Communications Psychology, 3, Article 183. https://doi.org/10.1038/s44271-025-00362-y
  4. Cankaya, O., Rohatyn-Martin, N., Leach, J., Taylor, K., y Bulut, O. (2023). Preschool children’s loose parts play and the relationship to cognitive development: A review of the literature. Journal of Intelligence, 11(8), 151. https://doi.org/10.3390/jintelligence11080151
  5. Chen, J. J. (2024). A scoping study on AI affordances in early childhood education. Journal of Artificial Intelligence Research, 81, 701–740. https://doi.org/10.1613/jair.1.16882
  6. Hu, H. (2025). Using action research to infuse nature-based loose parts play into the Kindergarten Program. Early Years, 45(3–4), 383–399. https://doi.org/10.1080/09575146.2024.2358422
  7. Ljungcrantz, L. (2026). The interaction of AI and early childhood education. A state-of-the-art review 2020–2024. Early Childhood Education Journal, 54, 3565–3581. https://doi.org/10.1007/s10643-025-02079-3
  8. Lyu, Y., y McNair, L. (2026). Bridging Froebel and AI: Reconceptualizing play pedagogy in Chinese context. Education Sciences, 16(3), 390. https://doi.org/10.3390/educsci16030390
  9. Miao, F., y Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO. https://doi.org/10.54675/EWZM9535
  10. NAEYC. (2022). Developmentally appropriate practice in early childhood programs serving children from birth through age 8 (4.ª ed.). NAEYC. https://www.naeyc.org/resources/pubs/books/dap-fourth-edition
  11. Nikolopoulou, K. (2025). Child-centered integration of generative AI in early learning: Balancing promises and challenges. AI, Brain and Child. https://doi.org/10.1007/s44436-025-00023-1
  12. OECD. (2021). Starting Strong VI: Supporting meaningful interactions in early childhood education and care. OECD Publishing. https://doi.org/10.1787/f47a06ae-en
  13. OECD. (2023). Empowering young children in the digital age (Starting Strong). OECD Publishing. https://doi.org/10.1787/50967622-en
  14. Su, J., y Yang, W. (2022). Artificial intelligence in early childhood education: A scoping review. Computers and Education: Artificial Intelligence, 3, 100049. https://doi.org/10.1016/j.caeai.2022.100049
  15. Su, J., y Zhong, Y. (2022). Artificial Intelligence (AI) in early childhood education: Curriculum design and future directions. Computers and Education: Artificial Intelligence, 3, 100072. https://doi.org/10.1016/j.caeai.2022.100072
  16. U.S. Department of Education, Office of Educational Technology. (2023). Artificial intelligence and the future of teaching and learning: Insights and recommendations. U.S. Department of Education. https://www.ed.gov/sites/ed/files/documents/ai-report/ai-report.pdf
  17. UNESCO. (2021). Recommendation on the ethics of artificial intelligence. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000381137
  18. Zeng, H. Q., y Ng, S. C. (2025). Free play matters: Promoting kindergarten children’s science learning using questioning strategies during loose parts play. Early Childhood Education Journal, 53(7), 2373–2388. https://doi.org/10.1007/s10643-024-01741-6