1. Introduction and problem

In the 3-to-6 age band —kindergarten, preschool, CENDI, early childhood school— a package of five artifacts has been installed as if it counted as support for sand play: a sensor or wearable that “scores construction accuracy” or a “sandcastle accuracy score” / “STEM sand score” / “molding accuracy score” for the child; a computer-vision / sandbox camera system that labels “mess” or “sand outside the tray” or triggers “mess / off-task sand spilling” alerts; a chatbot/GenAI that generates “sandcastle plans” or “sand play lesson plans” or “kinetic sand STEM activities” by prompt (for the child or as a recipe the adult reads); a dashboard of “minutes in the sandbox” / “construction score” for administrative surveillance; and an AR layer that “guides” molding or predicts “castle stability.” The five allow centers to display that they “already do sand play with AI.” The leap —from molding score, mess alert, generated plan, minutes metric or AR guidance to claiming support— is not authorized by AI-in-ECE mappings nor by the pedagogy of sand play when there is real sand (dry/wet, dig, pour, mold, build without expected product), physical properties of the granular material (flow, compaction, moldability), child agency, adult co-presence that names properties and sustains conversation, and tolerance of sand mess outside the tray as a condition of the medium, not as error.

The thesis of this article is restrictive. A sensor or wearable that “scores construction accuracy” or a “sandcastle accuracy score” / “STEM sand score” / “molding accuracy score” for the child, a computer-vision / sandbox camera system that labels “mess” or “sand outside the tray” or triggers “mess / off-task sand spilling” alerts, a chatbot/GenAI that generates “sandcastle plans” or “sand play lesson plans” or “kinetic sand STEM activities” by prompt, a dashboard of “minutes in the sandbox” / “construction score” for administrative surveillance, or an AR layer that “guides” molding or predicts “castle stability” do not constitute support for sand play in the early years. In early childhood sand play develops with real sand, child agency, adult co-presence that names properties (Iivonen et al., 2025; Sando et al., 2023), and tolerance of mess; not a molding score, a mess classifier, a plan generator that closes uncertainty, or a minutes dashboard. Iivonen et al. (2025) synthesize sand play (not clinical sandplay therapy) and outcomes ages 0–8: synthesis framework. Gutiérrez et al. (2022) anchor the sandbox as a didactic resource in Ecuadorian early education: craft of the sand medium, without converting this article into the generic motor-skills piece already published. Suryaningsih and Khoerunnisa (2024) explore kinetic sand and socialization/independence: finding with limits and marked transfer. Anggraeni et al. (2025) —N = 36 (18+18); ages 3–5— report effects on fine motor and cognitive development: marked finding; read via the kinetic-sand medium, not as a motor-skills axis. That authorizes asking what was measured: sandcastle accuracy, mess label, lesson plan by prompt, construction score or AR prediction —not the practice when a child digs, pours, molds, compares dry/wet and receives co-presence that names properties without closing uncertainty.

The problem is worsened by six category confusions. First: sand play is not water play —the axis is the granular (flow, compaction, moldability), not the already published liquid—. Second: it is not generic outdoor —the sandbox may be indoor; the axis is sand, not outdoor minutes—. Third: it is not loose parts —Sando et al. (2023) anchor objects-in-play; here the medium is sand, not a Nicholson collection—. Fourth: it is not motor skills as axis —Anggraeni et al. (2025) and Gutiérrez et al. (2022) may report lateral motor outcomes; the axis is the sand medium—. Fifth: it is not clinical sandplay therapy —Liu et al. (2023) is a peripheral PCST contrast ≠ classroom sand play—. Sixth: it is not free play as axis. This work does not recycle water play, outdoor, free play, loose parts, risky play, block play/spatial, symbolic play, motor skills, STEM/robotics, artistic creativity, SEL, executive functions, environmental education, scientific inquiry as axis, literacy, orality, UDL, documentation, formative assessment, participation, planning, joint attention, adult–child interactions as generic axis, teacher education, music or numeracy. Galbraith (2022) is a peripheral contrast of play pedagogies, not a teacher-education axis already published. The question is what counts as support for sand play when a center “does AI and sand play.”

There is, moreover, an economy of coordination: the five artifacts fit on a slide; the episode with real sand, agency, tolerated mess and co-presence that names properties does not. NAEYC (2022) and OECD (2021, 2023) require meaningful interactions and subordinated digitalization. Inference: support is not fulfilled by scoring construction or generating plans by prompt. The contributions are three: separate craft from artifacts; examine marked cases (Iivonen et al., 2025; Suryaningsih & Khoerunnisa, 2024; Anggraeni et al., 2025; Gutiérrez et al., 2022; Liu et al. and Galbraith peripheral); and offer four tests of support for sand play.

2. State of the art: from situated sand-play practice to the artifact on display

It is useful to separate four strata that the “AI for sand play in ECE” market often blends. The first is the construct of sand play ages 3–6 as open exploration of a granular material with real physical properties —dry vs wet sand, digging, pouring, molding, open construction, flow, compaction, moldability— and sand mess outside the tray as a condition of the medium (Iivonen et al., 2025; Gutiérrez et al., 2022; Suryaningsih & Khoerunnisa, 2024; Anggraeni et al., 2025; Sando et al., 2023). The second is the craft that cultivates it —real sand in sandbox, sand table or sand tray; agency without expected product; co-presence that names properties; tolerance of mess; not substituting with sandcastle accuracy or a mess classifier— (Iivonen et al., 2025; Gutiérrez et al., 2022; Galbraith, 2022, peripheral contrast; NAEYC, 2022; OECD, 2021, 2023). The third is evidence on AI affordances in ECE and child-centered GenAI —without equating them to situated sand play with real granular— (Chen, 2024; Su & Yang, 2022; Su & Zhong, 2022; Ljungcrantz, 2026; Nikolopoulou, 2025). The fourth is the rights, systems and developmentally appropriate practice framework that treats the 3–6-year-old as a subject who explores the granular, not as a vector of signals for a molding-accuracy pipeline nor as a passive recipient of GenAI plans or AR castle-stability predictions that close uncertainty (UNESCO, 2021; Miao & Holmes, 2023; U.S. Department of Education, 2023).

In construct and craft, Iivonen et al. (2025) articulate sand play and outcomes 0–8: synthesis framework; marked transfer when the range exceeds 3–6. Gutiérrez et al. (2022) situate the sandbox as a didactic resource: craft finding/framework. Suryaningsih and Khoerunnisa (2024) provide kinetic sand and socialization/independence: finding with limits. Anggraeni et al. (2025) —N = 36; ages 3–5— report effects on fine motor and cognitive development: marked finding; read via kinetic sand, not as a motor-skills article. Sando et al. (2023) —N = 928— associate objects (sand/water) with deep-level learning: objects-in-play, not loose parts. Liu et al. (2023) —PCST RCT; N = 52 dyads, 43 completed— is a peripheral clinical contrast. Galbraith (2022) contributes play pedagogies in preservice: peripheral adult contrast. Inference: a sandcastle accuracy score does not name granular properties; a child who digs and an adult who converses do.

In the artifact stratum, Chen (2024), Su and Yang (2022), Su and Zhong (2022) and Ljungcrantz (2026) map affordances, curricula and the state of the art of AI in ECE without equating them to situated sand play. Nikolopoulou (2025) balances promises and challenges of child-centered GenAI under teacher mediation: caution framework. 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 generators of “sandcastle plans” or kinetic sand STEM activities by prompt—. U.S. Department of Education (2023) requires AI to support, not replace, professional judgment. OECD (2021, 2023) anchor meaningful interactions and subordinated digitalization; 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 artifacts, but to articulate a pedagogical-category argument with verified sources. Inclusion criteria: (a) 2021–2026, with marked transfer when the sample is not equivalent to ages 3–6; (b) sand play, sandbox, sand table/tray, kinetic sand, granular properties, or AI in ECE with artifact/craft relevance; (c) kindergarten, preschool, CENDI or ages 3–6; (d) peer-reviewed journal, DOI or NAEYC/UNESCO/OECD report; (e) verifiable DOI or editorial page. Axes already used in this series were excluded as central objects —water play, generic outdoor, free play, loose parts, risky play, block play/spatial, symbolic play, STEM/robotics, artistic creativity, motor skills as axis, SEL, executive functions, environmental education, scientific inquiry as axis, literacy, orality, UDL, documentation, formative assessment, participation, planning, joint attention, adult–child interactions as generic axis, teacher education, music and numeracy—. Liu et al. (2023) only as peripheral clinical contrast; Galbraith (2022) only as peripheral contrast of play pedagogies. Gutiérrez et al. (2022) are read for the sandbox, not as a motor-skills axis. Sando et al. (2023) are read for objects-in-play (sand), not as loose parts or water play. Anggraeni et al. (2025) and Suryaningsih and Khoerunnisa (2024) are read for kinetic sand with marked limits.

The search was executed on 3 September 2026 (slot 17:02 America/Mexico_City) on DOI pages, Crossref, Springer, Elsevier, Wiley, Taylor & Francis, MDPI, JAIR, OECD iLibrary, UNESDOC, NAEYC and editorial sites. Each source was verified. Empirical finding, framework and pedagogical inference were distinguished. No N, d, r, AUC or DOI were invented: when an artifact lacks a verified study with sandcastle/molding-accuracy scoring, mess/off-task sand spilling alerts, construction-score dashboard or AR castle-stability prediction metrics in 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 play-pedagogies contrast (Galbraith, 2022). The eighteen verified corpus sources were used.

4. Case 1. Sensor/sandcastle accuracy score / STEM sand score / molding accuracy score, CV/sandbox camera with mess or off-task sand spilling alerts, GenAI of “sandcastle plans” / sand play lesson plans / kinetic sand STEM activities, minutes/construction-score dashboard or AR that guides molding or predicts stability do not constitute support for sand play

Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) saturate the portrait of the artifact ceiling when AI in ECE is presented as if it were support for sand 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 sandcastle-accuracy wearable cultivates open granular exploration with flow, compaction or moldability. Su and Yang (2022) review the AI-in-ECE field: synthesis finding on trends, not on tolerated mess or co-presence that names sand properties. Ljungcrantz (2026) reviews AI–ECE interaction 2020–2024: state-of-the-art finding, not situated sand play. Pedagogical inference, marked as such: the gesture “the child had high sandcastle accuracy = there was support for sand play” is a construction-precision analytics ceiling. Iivonen et al. (2025), Gutiérrez et al. (2022) and Sando et al. (2023) ask for real sand, conversation about properties and objects-in-play; a score useful for administration can coexist with absence of agency, tolerated mess and co-presence that names without closing.

Nikolopoulou (2025) and Miao and Holmes (2023) name the risk of the generator of “sandcastle plans” or “sand play lesson plans” / “kinetic sand STEM activities” from a prompt that closes uncertainty. Nikolopoulou (2025) balances promises and challenges of child-centered 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 lesson plans by prompt versus a sand-play episode with real sand and co-presence that sustains conversation (Iivonen et al., 2025; Gutiérrez et al., 2022). Inference: producing a “sandcastle” or “kinetic sand STEM” plan by prompt may be subordinated teacher preparation; support for sand play begins when there is real granular (dry/wet), child agency to dig, pour or mold without expected product, mess as a condition of the medium, and an adult who names properties without turning the episode into a recipe. Su and Zhong (2022) propose AI curriculum design in ECE as a future direction —AI literacy curriculum, not situated sand play—. Limit inference: an AI curriculum does not sign granular exploration.

The computer-vision / sandbox camera that labels “mess” or “sand outside the tray” or triggers “mess / off-task sand spilling” alerts, the “minutes in the sandbox” / “construction score” dashboard and the AR layer that “guides” molding or predicts “castle stability” lack, in the verified corpus, trials with reported N, d, r or AUC for sandcastle/molding-accuracy scoring, mess classification, construction-score dashboard or AR prediction in preschool ages 3–6; figures are not invented. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map analytics and generation as an AI-in-ECE trend, without equivalence to mediated sand play. Anggraeni et al. (2025) measure kinetic sand play with N = 36: situated practice with limit, not a mess classifier. Sando et al. (2023) measure human observations (N = 928): objects-in-play, not a dashboard. Galbraith (2022) documents play pedagogies: legitimate peripheral contrast only when the adult sustains the craft. Restrictive inference: Galbraith ≠ sandcastle-accuracy sensor; GenAI plans ≠ Iivonen/Gutiérrez who open without closing; teacher empowerment ≠ construction score; stability AR ≠ open construction. UNESCO (2021) and U.S. Department of Education (2023) require human oversight. The five artifacts share one substitution grammar: the score speaks for exploration; the alert substitutes tolerance of mess; GenAI closes; the dashboard substitutes observation; AR directs what craft leaves open.

It is useful to specify the ceiling without inventing effects. A sensor can score construction; Iivonen et al. (2025) articulate human sand play, not AUC. A camera can label “off-task sand spilling”; craft treats mess as a condition of the medium. A GenAI can print plans; Miao and Holmes (2023) require validation. An AR can predict stability; Gutiérrez et al. (2022) anchor the sandbox. A dashboard can count minutes; NAEYC (2022) and OECD (2021) anchor practice. Inference: “high sandcastle accuracy” or “lesson plan according to the model” does not sign support.

5. Case 2. What the kindergarten does when there is support for sand play: granular properties, Iivonen et al., Suryaningsih and Khoerunnisa, Anggraeni et al., Gutiérrez et al., Sando et al. —with Liu et al. and Galbraith as peripheral contrast

Iivonen et al. (2025) saturate the synthesis floor: mixed-methods systematic review on sand play and physical, cognitive and socioemotional outcomes ages 0–8. Status: synthesis framework / review finding. Marked transfer: range 0–8; read with caution toward ages 3–6, and pedagogical sand play is distinguished from clinical sandplay therapy. Pedagogical inference: a center may call it support for sand play when it protects real sand, agency and conversation about the granular —not when it scores construction accuracy—. Sand play is not a sandcastle accuracy score: it is digging, pouring, molding, comparing dry/wet, building without expected product, and sustaining mess as a condition of the medium, while the adult names flow, compaction or moldability.

Gutiérrez et al. (2022) anchor the sandbox as a didactic resource in Ecuadorian early education. Status: craft finding/framework of the sand medium. Marked distinction: although they discuss gross motor skills, they are cited here for the sandbox, not converting this article into the motor-skills piece already published. Inference: the object is the sand medium, not a STEM sand score. Suryaningsih and Khoerunnisa (2024) explore kinetic sand and preschool socialization/independence: empirical finding; sample/design limit and cautious transfer. Anggraeni et al. (2025) —N = 36 (18+18); ages 3–5; quasi-experimental— report effects on fine motor and cognitive development. Status: empirical finding. Marked limit: N = 36; one context; read via kinetic sand / sand play, not as a motor-skills axis nor as a mess-camera or GenAI trial. Inference: the measured object is mediated practice, not a molding accuracy score. Sando et al. (2023) —N = 928— anchor objects (including sand) in deep-level learning: sand as object-in-play, not loose parts or water play. NAEYC (2022) and OECD (2021, 2023) anchor meaningful interactions and subordinated digitalization.

The peripheral clinical contrast is Liu et al. (2023): Parent-Child Sandplay Therapy (PCST) RCT; N = 52 dyads, 43 completed. Status: finding of clinical sandplay THERAPY, not classroom sand play. Marked transfer: clinical ASD population; read only to fix therapy ≠ kindergarten sand play. Inference: PCST does not authorize the five artifacts as classroom sand play. The peripheral adult contrast is Galbraith (2022): play pedagogies in initial teacher education. Status: framework/finding toward the adult. Inference: subordinated digital —an adult who deepens human play pedagogies to sustain conversation at the sandbox— can be coherent; it does not authorize the five artifacts as craft. Galbraith is not converted into a teacher-education article already published. Mediation —NAEYC (2022), OECD (2021); Iivonen et al. (2025); Gutiérrez et al. (2022)— is co-presence that offers real sand, tolerates mess, names properties, observes agency without expected product and expands without a digital script that closes. Inference: pedagogical observation reads the episode with the granular; the dashboard counts minutes. A dashboard does not build lived moldability; a child and an adult who converse about dry/wet do.

6. Case 3. Sand play ≠ water play, generic outdoor, loose parts, motor skills as axis, clinical sandplay therapy or free play

The first frontier is already published water play: sand play shares with water play the logic of a material medium with real physical properties and tolerated mess, but the axis here is the granular (flow, compaction, moldability; dry/wet; dig, pour, mold), not the liquid of the water table. Inference: “we have a water table = there is sand play” does not sign; nor is a sandcastle-accuracy sensor water play or sand play. The second is generic outdoor/nature play: the sandbox may be outdoor or, centrally here, indoor sand table / sand tray; the axis is sand, not outdoor minutes. Inference: “we have a yard = there is sand play” does not sign. The third is loose parts: Sando et al. (2023) anchor objects (sand/water); here the medium is sand, not a Nicholson collection. The fourth is motor skills as curricular axis: Anggraeni et al. (2025) and Gutiérrez et al. (2022) may report motor outcomes; they do not authorize reducing sand play to motor curriculum or to “molding accuracy” tracking. The fifth is clinical sandplay therapy: Liu et al. (2023) fix PCST as contrast; therapy ≠ classroom sand play. The sixth is free play as axis; a GenAI of sandcastle plans closes uncertainty. UNESCO (2021), Miao and Holmes (2023) and U.S. Department of Education (2023) require human oversight. Inference: coherent AI use in ages 3–6 remains on the adult side (Galbraith, 2022), contrasted with Iivonen et al. (2025) and the sandbox (Gutiérrez et al., 2022). It does not enter as scorer, mess classifier, GenAI of plans, dashboard or predictive stability AR.

7. Inferential framework: four tests to affirm support for sand play, not an artifact

The framework that follows is pedagogical inference of this article, 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 does not pass them, it cannot declare that the five artifacts constitute support for sand play.

7.1. Test of situated practice with real sand —dry/wet, dig, pour, mold, open construction—, child agency and mess as a condition of the medium, not of the sandcastle/molding-accuracy sensor or the AR layer that guides or predicts stability. Iivonen et al. (2025), Gutiérrez et al. (2022), Suryaningsih and Khoerunnisa (2024), Anggraeni et al. (2025) and Sando et al. (2023) define the craft as observable granular exploration. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map analytics and generation as AI affordances, not as sand play. Inference: evidence of support for sand play is verified in whether the child explored real sand with agency and without expected product. If the center’s “evidence” is a sandcastle accuracy score or an AR castle-stability prediction, the center has done analytics or digital direction that closes, not support for sand play.

7.2. Test of adult co-presence that names properties and sustains conversation, not of GenAI “sandcastle plans” / sand play lesson plans / kinetic sand STEM activities. Iivonen et al. (2025), Gutiérrez et al. (2022), Galbraith (2022) and NAEYC (2022) situate conversation, sandbox and play pedagogies / developmentally appropriate practice. Miao and Holmes (2023) require pedagogical validation of generative AI. Inference: producing plans or STEM activities by prompt that close uncertainty does not demonstrate co-presence that names flow, compaction or moldability. A script may exist; it does not sign the sand-play episode.

7.3. Test of human mediation —including peripheral contrasts Liu et al. and Galbraith—, not of the minutes/construction-score dashboard or the CV/camera of mess or off-task sand spilling alerts. OECD (2021, 2023) require meaningful interactions and subordinated digitalization. Galbraith (2022) anchors play pedagogies toward the adult. Liu et al. (2023) fix the therapy ≠ classroom frontier. Sando et al. (2023) anchor objects-in-play and deep-level learning observed by humans. Inference: “high minutes in the sandbox” or a “mess alert” may raise an administrative threshold without raising co-presence quality or child exploration. Pedagogical observation reads the episode with the granular; the dashboard counts surveillance. Human adult knowledge/pedagogy is legitimate periphery; the wearable/CV that substitutes exploration or mess tolerance is not.

7.4. Test of category distinction and professional judgment, not of the product catalog. Sand play ≠ water play, generic outdoor, loose parts, motor skills as axis, clinical sandplay therapy or free play. Sando et al. (2023) prevent reducing sand-objects to loose parts as axis. Iivonen et al. (2025) and Liu et al. (2023) prevent confusing pedagogical sand play with sandplay therapy. Anggraeni et al. (2025) and Gutiérrez et al. (2022) prevent confusing lateral motor outcomes with motor curriculum as axis or with a STEM sand 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 judgment. Inference: a center cannot treat the child as a sandcastle-accuracy vector nor as exclusive recipient of GenAI plans. Support for sand play is not fulfilled by better scoring “accuracy” or surveilling “mess.” It is fulfilled by practicing real sand, child agency to explore without expected product, adult co-presence that names properties and sustains conversation, and tolerance of mess.

The framework admits digital subordinated to adult pedagogies/preparation and validated human mediation (Galbraith, 2022; Chen, 2024; Su & Yang, 2022; Ljungcrantz, 2026). It rejects declaring support via the five artifacts (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 the five artifacts and exercising support for sand play. Iivonen et al. (2025), Gutiérrez et al. (2022), Suryaningsih and Khoerunnisa (2024), Anggraeni et al. (N = 36), Sando et al. (N = 928), Liu et al. and Galbraith (peripheral) sustain the craft or its frontier; Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map AI without equivalence to situated sand play. It is not a finding that the artifacts produce the craft.

The second is between automated assessment of sandcastle/molding accuracy or mess and pedagogy of the granular medium. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) objectivize analytics; they do not report verified trials of scoring or mess alerts in preschool ages 3–6 —metrics are not invented—. Inference: declaring support via a construction score or sand-outside-tray alert is inverted pedagogy.

The third is between GenAI of plans / stability AR and integral craft. Miao and Holmes (2023) and Nikolopoulou (2025) set limits; Iivonen et al. (2025) and Gutiérrez et al. (2022) fix situated sand play and sandbox. Inference: selling plans by prompt or AR as “sand play with AI” confuses product with granular exploration. Anggraeni et al. (2025) confirm practice N = 36 with limit without authorizing a score; UNESCO (2021), Miao and Holmes (2023) and U.S. Department of Education (2023) subordinate AI to professional judgment.

The four tests in section 7 read these tensions. The empirical contrast —Iivonen; Gutiérrez; kinetic sand with limits; N = 36; N = 928; peripheral PCST and Galbraith— defines the floor the artifacts do not reach alone. Inference: support erodes when they are treated as if they were the practice. Subordinated digital is legitimate toward the adult (Galbraith, 2022); inverting the sequence is not (NAEYC, 2022; OECD, 2021, 2023; Iivonen et al., 2025).

9. Limits

This review is narrative. It does not apply its own PRISMA nor estimate combined effects. Iivonen et al. (2025) synthesize ages 0–8: partial transfer toward 3–6. Suryaningsih and Khoerunnisa (2024) provide kinetic sand with limits. Anggraeni et al. (2025) measure N = 36: marked limit. Gutiérrez et al. (2022) anchor the sandbox read via the sand medium, not as a motor-skills axis. Sando et al. (2023; N = 928), Liu et al. (2023; peripheral PCST) and Galbraith (2022; peripheral play pedagogies) delimit frontiers. Chen (2024), Su and Yang (2022), Su and Zhong (2022), Ljungcrantz (2026) and Nikolopoulou (2025) map AI in ECE, not sandcastle accuracy or mess alerts. No verified trials of the five artifacts in CENDI ages 3–6 were located; they are discussed as a category ceiling. NAEYC, UNESCO and OECD are framework sources. Section 7 inferences are category hypotheses, not implementation evidence.

10. Conclusions

A sensor or wearable that “scores construction accuracy” or a “sandcastle accuracy score” / “STEM sand score” / “molding accuracy score” for the child, a computer-vision / sandbox camera system that labels “mess” or “sand outside the tray” or triggers “mess / off-task sand spilling” alerts, a chatbot/GenAI that generates “sandcastle plans” or “sand play lesson plans” or “kinetic sand STEM activities” by prompt, a dashboard of “minutes in the sandbox” / “construction score” for administrative surveillance, or an AR layer that “guides” molding or predicts “castle stability” do not constitute support for sand play in an early childhood center. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) confirm AI affordances without equivalence to situated sand play. Nikolopoulou (2025) and Miao and Holmes (2023) set GenAI limits. Liu et al. (2023) require reading PCST as peripheral contrast —therapy ≠ classroom sand play—. Galbraith (2022) requires reading play pedagogies as peripheral contrast —adult ≠ sensor, mess CV, GenAI, dashboard or AR—. Iivonen et al. (2025) and Gutiérrez et al. (2022) fix sand play and sandbox. When there is support, there is situated practice: systematic review of sand play (Iivonen et al., 2025); sandbox as resource (Gutiérrez et al., 2022); kinetic sand with limits (Suryaningsih & Khoerunnisa, 2024; Anggraeni et al., 2025, N = 36); objects-in-play N = 928 (Sando et al., 2023); DAP and interactions (NAEYC, 2022; OECD, 2021, 2023). Sand play is distinguished from water play, generic outdoor, loose parts, motor skills as axis, clinical sandplay therapy and free play. Guidance requires 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, PCST or play pedagogies toward the adult, it does not translate them into support via score or mess alert. Accompanying children aged three to six in sand play is to practice real sand (dry/wet, dig, pour, mold, build without expected product), child agency, adult co-presence that names properties, and tolerance of mess as a condition of the medium. Anything else is a sandcastle/molding-accuracy sensor, a mess classifier, a plan generator that closes uncertainty, a minutes dashboard and an AR layer that guides or predicts stability. It is not support for sand 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. Anggraeni, R., Mariani, D., y Rosuliana, N. E. (2025). The effect of the kinetic sand play method on fine motor and cognitive development in preschool children. Journal of Sports Nursing, Medical, And Health, 1(1), 1–9. https://doi.org/10.69606/sportnursmedhealth.v1i01.289
  2. Chen, J. J. (2024). A scoping study on AI affordances in early childhood education: Mapping the global landscape, identifying research gaps, and charting future research directions. Journal of Artificial Intelligence Research, 81, 701–740. https://doi.org/10.1613/jair.1.16882
  3. Galbraith, J. (2022). “A prescription for play”: Developing early childhood preservice teachers’ pedagogies of play. Journal of Early Childhood Teacher Education, 43(3), 474–494. https://doi.org/10.1080/10901027.2022.2054035
  4. Gutiérrez Tasinchana, V. P., Arroba López, G. A., Ballesteros Casco, T. Y., y Sánchez Fernández, I. (2022). El arenero: un recurso didáctico para el desarrollo de la motricidad gruesa en la educación inicial. ConcienciaDigital, 5(1.1), 179–210. https://doi.org/10.33262/concienciadigital.v5i1.1.1993
  5. Iivonen, S., Kettukangas, T., Soini, A., y Viholainen, H. (2025). Sand play and 0- to 8-year-old children’s physical, cognitive and socioemotional outcomes: A mixed-methods systematic review. Child: Care, Health and Development, 51(1), Article e70034. https://doi.org/10.1111/cch.70034
  6. Liu, G., Chen, Y., Ou, P., Huang, L., Qian, Q., Wang, Y., He, H.-G., y Hu, R. (2023). Effects of Parent-Child Sandplay Therapy for preschool children with autism spectrum disorder and their mothers: A randomized controlled trial. Journal of Pediatric Nursing, 71, 6–13. https://doi.org/10.1016/j.pedn.2023.02.006
  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(5), 3565–3581. https://doi.org/10.1007/s10643-025-02079-3
  8. Miao, F., y Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO. https://doi.org/10.54675/EWZM9535
  9. 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
  10. Nikolopoulou, K. (2025). Child-centered integration of generative AI in early learning: Balancing promises and challenges. AI, Brain and Child, 1(1). https://doi.org/10.1007/s44436-025-00023-1
  11. OECD. (2021). Starting Strong VI: Supporting meaningful interactions in early childhood education and care. OECD Publishing. https://doi.org/10.1787/f47a06ae-en
  12. OECD. (2023). Empowering young children in the digital age (Starting Strong). OECD Publishing. https://doi.org/10.1787/50967622-en
  13. Sando, O. J., Sandseter, E. B. H., y Brussoni, M. (2023). The role of play and objects in children’s deep-level learning in early childhood education. Education Sciences, 13(7), 701. https://doi.org/10.3390/educsci13070701
  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. Suryaningsih, C., y Khoerunnisa, D. (2024). Exploring the impact of kinetic play sand on the socialization and independence of preschool children. Asian Journal of Healthy and Science, 3(2), 86–92. https://doi.org/10.58631/ajhs.v3i2.99
  17. 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
  18. UNESCO. (2021). Recommendation on the ethics of artificial intelligence. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000381137