1. Introduction and problem

In the 3-to-6-year span—kindergarten, preschool, CENDI, early childhood school—a package of three artifacts has been installed as if it counted as music education. The first is a pitch-correction app that marks the child’s singing as “correct/incorrect”: a classification or regression pipeline that compares fundamental frequency with a pattern and declares that the child “already sang in tune.” The second is a song or lullaby generator from a prompt: a model that produces audio or lyrics under textual instruction and presents the track as if it were a pedagogical musical experience. The third is an affective-computing system that classifies “musical engagement” from video or audio: a multimodal framework that labels emotion or attention and delivers the dashboard as if it were practice of listening, singing and rhythm. All three are visible and cheap in coordination time. They allow a center to exhibit that it “already does music education with artificial intelligence.” The leap—from obtaining a pitch score, receiving a generated lullaby or accumulating engagement labels to asserting that music education exists—is authorized neither by vocal-scoring or affective-computing evidence nor by what music pedagogy measures when there is situated practice of shared singing, rhythmic-body play, musical movement, timbral exploration and adult mediation in routines and circle.

The thesis of this article is restrictive. A pitch-correction app that marks a child’s singing as “correct/incorrect,” a song or lullaby generator from a prompt, or an affective-computing system that classifies “musical engagement” from video or audio do not constitute music education in early childhood education. At this stage music education is situated practice of shared singing, rhythmic-body play, active listening and adult mediation—musical movement and timbral exploration cultivated in routines, music circle and games with adults who model, invite and extend—; not a pitch score, an emotion dashboard or a generated track that replaces the group’s voice. Li (2026) reports a CNN-BiLSTM pitch-correction system in children’s music teaching with N = 40 children, pitch accuracy of 91.6% (+13% versus control) and reduction of deviation from 18.5 to 9.2 Hz: a finding of tuning accuracy, not of kindergarten musical craft. Shi (2026) proposes an affective-computing framework with MoE architecture and neural radiance fields in preschool music education (N = 120; 60 multimodal hours; emotional accuracy 89.7%; engagement +30%): a finding of affective classification, not of mediated shared singing. That does not authorize translating “pitch correction is installed” or “an engagement classifier is running” as “music education exists.” It authorizes asking what was measured: often Hz of deviation or emotional accuracy, not the reiterated practice when a four-year-old sings with others, marks the pulse with the body, listens to timbres and participates in a circle with a mediating adult.

The problem is aggravated by five category confusions that product sheets do not mention and that this article separates rigorously. First: music education is not visual artistic creativity—a painting workshop, collage or plastic “expression” sold as “art with AI.” Second: it is not motor development or fundamental movement skills (FMS)—although musical movement activates the body, the construct is not exhausted by locomotion or object control. Third: it is not STEM or educational robotics. Fourth: it is not SEL as a generic socio-emotional competencies program. Fifth: it is not orality or narration—although song and language touch, musical craft is not reduced to storytelling. This work does not recycle articles on motor development, artistic creativity, interactions, executive functions, SEL, STEM, free play, documentation, numeracy, orality, formative assessment, UDL, participation, planning or teacher training.

There is, moreover, a coordination economy that explains why the package sells so quickly. A pitch score, a generated lullaby and an engagement dashboard are legible in a supervision meeting: they fit on a slide, can be photographed, can be compared across rooms.

The contributions are three: to reconstruct the state of the art that separates music education (shared singing; rhythmic-body play; musical movement; timbral exploration; mediation in routines and circle) from pitch scoring, the prompt lullaby and the engagement dashboard; to examine three families of empirical cases; and to offer four tests for deciding when a kindergarten may assert that music education exists, and not merely a pitch-correction app, a generator or an affective classifier.

2. State of the art: from situated musical practice to the exhibited artifact

It is useful to separate four strata that the market for “AI for music education in early childhood” usually mixes. The first is the construct of music education for ages 3 to 6 as practice of singing, rhythm-body, listening and timbral exploration in everyday contexts of routine and circle (del Barrio and Arús, 2024; Şenol and Karaca, 2025; Webster and Holmes, 2026; NAEYC, 2022; OECD, 2021). The second is the pedagogical craft that cultivates it—shared singing, rhythmic-body play, musical movement, teacher mediation in routines and music circle, active listening— (del Barrio and Arús, 2024; Webster and Holmes, 2026; NAEYC, 2022; OECD, 2021, 2023). The third is the evidence on pitch correction, affective computing of engagement, digital platforms with AI and mappings of AI in ECE (Li, 2026; Shi, 2026; Roldan-Cardona et al., 2025; Chen, 2024; Su and Yang, 2022; Ljungcrantz, 2026; Nikolopoulou, 2025). The fourth is the framework of rights, systems and developmentally appropriate practice, which treats the 3–6-year-old as a subject of situated musical practice, not as a vector of Hz for a scoring pipeline nor as an emitter of affective signals for a dashboard (UNESCO, 2021; Miao and Holmes, 2023; European Commission, 2022; U.S. Department of Education, 2023).

In the construct and craft stratum, del Barrio and Arús (2024) systematically review 29 articles (2013–2023) on music and movement pedagogy in basic education and include cited evidence on programs that articulate rhythm, body and mediation—among them RAMSR/Williams et al. (2023) as evidence referred to in the review—. Status: synthesis finding on music+movement as a pedagogical field, not on AI. Pedagogical inference, marked as such: the craft a kindergarten may call music education is bodily-musical and mediated; not a pitch score. Şenol and Karaca (2025) investigate the effect of a music education programme on preschoolers’ motor creativity skills: a finding that a musical programme may be associated with motor creativity; not a finding that a pitch app does so. Webster and Holmes (2026) reconceptualize music as a central—not ancillary—pedagogical tool in an early years case study: small sample; craft framework. NAEYC (2022) requires developmentally appropriate practice and includes everyday musical practices—music and songs as routine and community—as part of the learning environment; OECD (2021) anchors ECEC quality in meaningful everyday interactions; OECD (2023) requires that digitalization empower without substituting those interactions. Inference: a pitch score does not sing with the child. An adult who organizes the circle, models a rhythmic motif, invites exploration of a timbre and listens to the group’s voice does.

In the artifact stratum, Li (2026) reports CNN-BiLSTM for pitch correction in children’s music teaching (N = 40; pitch accuracy 91.6%, +13% vs control; deviation 18.5→9.2 Hz). Status: empirical finding of tuning accuracy. Restrictive inference: pitch score ≠ situated music education. Shi (2026) reports MoE + NeRF for affective computing in preschool music education (120 preschool; 60 multimodal hours; emotional accuracy 89.7%; engagement +30%). Status: multimodal affective classification finding. Inference: classifying emotion ≠ mediated musical practice. Roldan-Cardona et al. (2025) examine music education with AI for inclusive and sustainable early childhood learning (N = 15; ages 3–6; Genially/Educaplay/Wordwall): motivation and pronunciation; no significant differences between groups; limited pedagogical transfer. Chen (2024), Su and Yang (2022), Ljungcrantz (2026) and Nikolopoulou (2025) map affordances, reviews and promises/challenges of AI in ECE without equating them to shared singing or the music circle.

In the systems stratum, UNESCO (2021) requires human oversight and AI ethics. Miao and Holmes (2023) set pedagogical validation and age thresholds for generative AI—a direct framework for the prompt-based lullaby generator—.

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 pitch correction or affective computing, but to articulate a pedagogical-category argument with verified sources. Inclusion criteria: (a) 2021–2026, with transfer explicitly marked when the sample or level does not equate to kindergarten ages 3–6; (b) music education, singing, rhythm, musical movement, pitch correction, affective computing, musical generation or AI in early childhood education; (c) relevance for kindergarten, preschool, CENDI or ages 3–6; (d) peer-reviewed journal, DOI or report from NAEYC, UNESCO, OECD, European Commission or education department; (e) verifiable DOI or editorial page. Excluded as central object were axes already used in this series—motor development/FMS, visual artistic creativity, CLASS/serve-and-return interactions, pedagogical documentation, SEL as substitute programme, executive functions, STEM/robotics, generic free play, numeracy, orality/narration, generative tutoring, gaps, privacy, UDL, academic formative assessment, family–school, continuous training, participation and planning—although some appear as category limits.

The search was executed on 28 August 2026 (slot 13:02 America/Mexico_City) on DOI pages, Crossref, Springer, Frontiers, Taylor & Francis, Emerging Science Journal, JAIR, OECD iLibrary, UNESDOC, NAEYC and editorial sites. Each source was verified against at least one of those pages. Empirical finding, conceptual or normative framework, and pedagogical inference marked as such were distinguished. Priority was given to the distinction music education / visual creativity / motor-FMS / STEM / SEL / orality, and to the caution of not translating pitch accuracy, emotional accuracy or dashboard-reported engagement into pedagogy of shared singing and the music circle.

4. Case 1. Pitch correction, lullaby generators or affective computing of engagement do not constitute music education

Li (2026) publishes in Frontiers in Artificial Intelligence and Applications the comparison that best names the ceiling of the first artifact when it is presented as music education. The study reports a CNN-BiLSTM pitch-correction system applied to children’s music teaching: N = 40 children; pitch accuracy of 91.6% with a +13% improvement versus control; pitch deviation reduced from 18.5 to 9.2 Hz. Status of the evidence. Empirical finding of accuracy and reduction of pitch error. Pedagogical inference, marked as such: this is the gesture a center copies when it “does music with AI that scores singing.” One records, estimates frequency, declares “correct/incorrect.” What exists is a technical ceiling of vocal scoring. Music education, in del Barrio and Arús (2024), Webster and Holmes (2026) and NAEYC (2022), asks for shared singing, rhythm-body and mediation in routines; not an Hz closer to the pattern. A child may “tune better” according to the model and, at the same time, not have sung with others, not have explored a timbre nor participated in a circle where an adult mediates mutual listening.

Shi (2026) saturates the portrait of the second and third artifacts when engagement is classified as if it were practice. In Discover Artificial Intelligence the author proposes an affective-computing framework for preschool music education with Mixture-of-Experts architecture and neural radiance fields: 120 preschool children; 60 hours of multimodal data; emotional accuracy 89.7%; engagement reported with a +30% increase. Status: finding of affective classification and engagement metrics in a multimodal framework. It is not a finding that installing an emotion classifier develops shared singing, rhythmic-body play or timbral exploration in classroom life. Inference, marked as such: classifying emotion or engagement authorizes saying that there is more reproducible measurement of affective signals; it does not authorize declaring “music education already exists.” The dashboard may rise and the circle may empty of adult voices that sing with the group.

Roldan-Cardona, Chacón-Castro, Jadán-Guerrero, Salvador-Ullauri and Acosta-Vargas (2025) name the face of digital platforms with AI when they are treated as inclusive and sustainable music education. In Emerging Science Journal they work with N = 15 children aged 3 to 6 using Genially, Educaplay and Wordwall; they report effects on motivation and pronunciation; they find no significant differences between groups; pedagogical transfer is limited. Status: empirical finding of a small sample in the kindergarten age range, with mixed results and without automatic equivalence to musical craft. Inference: a gamified platform may motivate or support pronunciation; it does not sign the music circle or rhythmic-body play. The song or lullaby generator from a prompt—a generative-AI artifact that Miao and Holmes (2023) place under pedagogical validation and age thresholds—inherits the same category risk: producing a track is not organizing shared singing. A generated lullaby may be sonic material; music education begins when there are voices that listen to one another, bodies that mark the pulse and an adult who mediates reiteration. Chen (2024), Su and Yang (2022), Ljungcrantz (2026) and Nikolopoulou (2025) confirm growth of AI in ECE without evidence that pitch correction, prompt lullabies or engagement classifiers substitute craft.

The three artifacts share the same grammar of substitution. Pitch correction substitutes pedagogical listening to singing with a binary verdict.

5. Case 2. What the kindergarten does when music education exists: shared singing, rhythm-body and mediation

del Barrio and Arús (2024) situate music and movement pedagogy as a reviewable empirical field: 29 articles between 2013 and 2023; articulation of rhythm, body and mediation; inclusion of cited evidence on programmes such as RAMSR/Williams et al. (2023) within the review. Status: synthesis finding on music+movement in basic education, with transfer marked when the level is not exclusively kindergarten ages 3–6, but with a framework directly usable for the bodily-musical craft of early childhood. Pedagogical inference, marked as such: this is the object an early childhood center may call music education. Music is not a pitch score: it is a growing domain of singing together, marking pulse, moving with musical intention, listening and exploring timbres in meaningful contexts. An app that marks “correct” without shared practice may celebrate an Hz and, at the same time, empty the circle of mediation. An engagement classifier may label “high attention” in video and not have captured whether the adult modeled the motif, adjusted the group’s tempo or invited transfer of the pattern to a clapping game.

Şenol and Karaca (2025) saturate the craft from a music education programme for preschoolers and its effect on motor creativity: a finding that the musical programme is associated with motor creativity skills. Status: empirical finding of a musical programme (not of a pitch app). Inference: when music education exists in early childhood, there is programme and mediated practice, not a feed of tuning alerts. Webster and Holmes (2026) reconceptualize music as a central pedagogical medium in an early years case: small sample; the value lies in situating music at the center of everyday work, not as ancillary to a dashboard. NAEYC (2022) anchors everyday musical practices—songs and music as routine and community—in developmentally appropriate practice. OECD (2021) anchors quality in everyday interactions; OECD (2023) requires subordinated digitalization. Inference: teacher mediation in routines and circle means noticing the quality of mutual listening, the invitation to sing, the group’s emotional safety and the opportunity to reiterate motifs; pedagogical observation of musical practice is professional reading, not a log of Hz or emotion labels. A dashboard does not model an ostinato. An adult who organizes the circle, sings with the child and observes transfer among entry routine, rhythmic play and farewell does.

Timbral exploration and active listening do not deny play: they orient it musically. del Barrio and Arús (2024) underline the music–movement link; Webster and Holmes (2026) underline pedagogical centrality.

It is useful to specify what counts as mediation in early childhood musical craft, without inventing data the sources do not report. Mediation, in the sense of NAEYC (2022) and OECD (2021), is adult presence that organizes, models, invites and extends: the adult initiates or sustains a routine song; offers a clapping pattern and waits for the group’s response; brings an instrument or sounding object and names the timbre; adjusts tempo when the group disorganizes; protects listening silence before a new motif.

6. Case 3. Music education is not visual creativity, nor motor-FMS, nor STEM, nor SEL, nor orality

The first category frontier is visual artistic creativity. A painting workshop, a collage or a “generated work” may coexist in a center; they do not constitute, by themselves, music education of shared singing and rhythm-body at ages 3–6. Status: pedagogical inference marked as such, anchored in del Barrio and Arús (2024), Webster and Holmes (2026) and NAEYC (2022) as musical construct. This article does not make visual creativity its central object: it names it to forbid the equivalence. A pitch-correction system is also not “digital musical art”: it is frequency estimation. Confusing “we have vocal scoring” with “there is mediated singing practice” is the category error the market exploits.

The second frontier is motor development / FMS. Şenol and Karaca (2025) show that a music education programme may be associated with motor creativity: a finding of empirical overlap, not of identity of constructs. Inference: musical movement activates the body; it does not turn music education into assessment of locomotion or object control. This article does not recycle the motor-development piece: it uses it only as a limit. The third frontier is STEM/robotics. A kit, a programming sequence or a metaphor of “sounding like a machine” do not constitute music education. The fourth frontier is generic SEL. Socio-emotional competencies may appear in the circle; they are not shared singing or timbral exploration. Restrictive inference: an engagement dashboard that rewards “positive emotion” does not turn affective reinforcement into musical craft. The fifth frontier is orality/narration. Song touches language; it is not reduced to telling stories. Inference: when a product promises “it improves narration because it generates lullabies,” another construct has been crossed. Chen (2024), Su and Yang (2022), Ljungcrantz (2026) and Nikolopoulou (2025) map AI in ECE; Miao and Holmes (2023) set limits for generative AI. Inference: the only AI use coherent with ages 3–6 remains on the side of the adult who organizes shared singing, rhythmic play and circle—as support for repertoire preparation or deferred professional feedback—subject to pedagogical validation. It does not enter as an autonomous “correct/incorrect” scorer, a generator that replaces the group’s voice or a classifier that declares musical engagement fulfilled.

7. Inferential framework: four tests to assert that music education exists, 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 a pitch-correction app that marks “correct/incorrect,” a song/lullaby generator from a prompt or an affective-computing system that classifies “musical engagement” constitute music education.

7.1. Test of situated practice of shared singing, rhythm-body and listening, not of the pitch score. del Barrio and Arús (2024) and Webster and Holmes (2026) define the craft as mediated bodily-musical practice. Li (2026) estimates pitch with CNN-BiLSTM (N = 40; 91.6% accuracy; 18.5→9.2 Hz). Inference: evidence of music education is verified in whether the child practiced singing, rhythm and listening in context. If the center’s “evidence” is an Hz dashboard or a “correct/incorrect,” the center has done scoring, not music education.

7.2. Test of the circle, routines and rhythmic-body play, not of the lullaby generator. NAEYC (2022) situates music and songs as routine and community. Miao and Holmes (2023) require pedagogical validation of generative AI. Inference: producing a lullaby by prompt does not demonstrate that there was a mediated music circle or reiterated rhythmic play. A track may sound; it does not sign the practice.

7.3. Test of teacher mediation and pedagogical observation of musical practice, not of the engagement dashboard. OECD (2021, 2023) require meaningful interactions and subordinated digitalization. Shi (2026) classifies emotion/engagement (120 preschool; 89.7% accuracy; +30% engagement). Inference: “high engagement” may raise a threshold without raising the quality of mediation. Pedagogical observation reads singing and rhythmic play; the dashboard counts a label.

7.4. Test of category distinction and professional judgment, not of the product catalogue. Music education ≠ visual creativity, motor-FMS, STEM, generic SEL or orality. Şenol and Karaca (2025) prevent confusing overlap with motor creativity with identity of constructs. Roldan-Cardona et al. (2025) show limits of digital platforms with N = 15. UNESCO (2021), Miao and Holmes (2023), the European Commission (2022), the U.S. Department of Education (2023), NAEYC (2022) and OECD (2021, 2023) require human oversight, pedagogical validation and not substituting professional judgment. Inference: a center cannot treat the child as a pitch vector or as an emitter of emotion labels. Music education is not fulfilled by scoring the singing audio better. It is fulfilled by practicing shared singing, rhythmic-body play, musical movement and timbral exploration with adults who mediate in routines and circle.

The framework admits the digital when it is subordinated to teacher repertoire preparation and validated professional observation (Roldan-Cardona et al., 2025, as auxiliary platform with limits; Li, 2026, and Shi, 2026, as subordinated scoring/classification, not as declared education). It rejects declaring music education by autonomous pitch correction, prompt lullaby or affective computing of engagement (Li, 2026; Shi, 2026; Miao and Holmes, 2023).

The four tests are read together. Passing only the first—there was singing—without mediation or category distinction does not suffice.

8. Discussion

Three tensions organize the discussion. The first is between exhibiting an artifact of musical measurement or generation and exercising music education. It is a finding that CNN-BiLSTM improves pitch accuracy and reduces Hz deviation (Li, 2026); that MoE + NeRF classify emotion and report engagement (Shi, 2026); and that Genially/Educaplay/Wordwall platforms with N = 15 show no significant differences between groups (Roldan-Cardona et al., 2025). It is a framework that music education at ages 3–6 is played out in shared singing, music+movement and everyday practices (del Barrio and Arús, 2024; Webster and Holmes, 2026; Şenol and Karaca, 2025; NAEYC, 2022). It is not a finding that score, generator or dashboard produce the craft the circle requires. The three artifacts measure, generate or label what engineering knows how to deliver and declare what only mediated practice would authorize.

The second is between automated assessment of singing/affect and everyday pedagogy. Li (2026) objectifies tuning; Shi (2026) objectifies engagement. Inference: insisting that the kindergarten “already has music education” because the model classifies enough correct tones or emotions is inverted pedagogy. An evaluative proxy is made to stand for practice of shared singing, rhythmic-body play and listening. The child becomes an emitter of Hz or affective signals; the adult, a supervisor of the classifier.

The third is between generated lullaby and integral musical craft. Miao and Holmes (2023) set limits for generative AI; Nikolopoulou (2025) balances child-centered promises and challenges. Inference: fragmenting music education into a track produced by prompt and selling it as “music education with AI” confuses sonic product with situated practice. The only AI use coherent with ages 3–6 remains on the side of the adult who organizes the circle and observes musical practice, subject to UNESCO, the European Commission and the U.S. Department of Education. An autonomous scorer or a generator because the system needs a product is not that use.

Additional pedagogical inference, marked as such: the artifact series—pitch correction that scores correct/incorrect, lullaby generator and affective computing of engagement—shares the same economy of visibility. Each produces an output legible for coordination: an Hz number, a track, an emotion label. OECD (2021) and NAEYC (2022) require, instead, that development of musical practice be verified in everyday classroom life. Chen (2024), Su and Yang (2022), Ljungcrantz (2026) and Nikolopoulou (2025) show growth of the AI-in-early-childhood field; that growth does not authorize the equivalence. Miao and Holmes (2023), UNESCO (2021), the European Commission (2022) and the U.S. Department of Education (2023) agree on subordinating AI to professional judgment.

A cross-reading of the cases reinforces the restrictive thesis without inventing effects the sources do not report. Li (2026) shows that a pitch score can be improved; it does not show that the kindergarten exercised shared singing.

9. Limits

This review is narrative. It does not apply PRISMA nor estimate primary combined effects. Li (2026) validates pitch correction (N = 40), not everyday pedagogy of the music circle. Shi (2026) validates multimodal affective computing (N = 120; 60 h), not a trial of mediated shared singing versus an autonomous dashboard. Roldan-Cardona et al. (2025) are N = 15 with digital platforms: small sample, no significant differences between groups, limited pedagogical transfer. del Barrio and Arús (2024) review 29 music+movement articles in basic education: transfer marked when the level is not exclusively ages 3–6. Şenol and Karaca (2025) associate a musical programme with motor creativity, not with AI. Webster and Holmes (2026) are a small-sample case study. Chen (2024), Su and Yang (2022), Ljungcrantz (2026) and Nikolopoulou (2025) map AI in ECE, not circle music education as primary variable. NAEYC, UNESCO and OECD are framework sources. No Latin American AI trials were located comparing adult-mediated shared singing versus autonomous pitch correction or affective computing of engagement in CENDI settings. The inferences in section 7 are pedagogical-category hypotheses, not implementation evidence.

10. Conclusions

A pitch-correction app that marks a child’s singing as “correct/incorrect,” a song or lullaby generator from a prompt, or an affective-computing system that classifies “musical engagement” from video or audio do not constitute music education in an early childhood education setting. Verified evidence does not authorize that declaration. Li (2026) shows pitch accuracy of 91.6% (+13% vs control) and reduction of deviation from 18.5 to 9.2 Hz with CNN-BiLSTM in N = 40—pitch scoring, not kindergarten craft—. Shi (2026) classifies emotion with 89.7% accuracy and reports +30% engagement in 120 preschool children and 60 multimodal hours—affective classification, not shared singing—. Roldan-Cardona et al. (2025) work with N = 15 (ages 3–6) on digital platforms without significant differences between groups and with limited pedagogical transfer. By contrast, when music education exists in early childhood, there is situated practice: music+movement reviewed in 29 articles (del Barrio and Arús, 2024); a musical programme associated with motor creativity in preschool (Şenol and Karaca, 2025); music as a central pedagogical medium in early years (Webster and Holmes, 2026); everyday musical practices and developmentally appropriate practice (NAEYC, 2022; OECD, 2021, 2023). Music education is distinguished from visual artistic creativity, from motor development/FMS, from STEM/robotics, from generic SEL and from orality/narration. Current guidance requires human oversight, pedagogical validation and not substituting professional judgment (UNESCO, 2021; Miao and Holmes, 2023; European Commission, 2022; U.S. Department of Education, 2023; Chen, 2024; Su and Yang, 2022; Ljungcrantz, 2026; Nikolopoulou, 2025).

Where sources do not measure a kindergarten, this article does not assert it. Where they measure pitch correction, affective computing or digital platforms, it does not translate them into pedagogical music education. Accompanying girls and boys aged three to six in music is to exercise practice of shared singing, rhythmic-body play, musical movement, active listening and timbral exploration, with teacher mediation in routines and music circle. The rest is pitch score, prompt lullaby and algorithmic classification of engagement. It is not music education in early childhood education, and it must not be presented as what it is not.

Laboratorio Editorial de NEXTECH.IA / Ingeniero Mitre.

References

  1. 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
  2. del Barrio, L., y Arús, M. E. (2024). Music and movement pedagogy in basic education: A systematic review. Frontiers in Education, 9, 1403745. https://doi.org/10.3389/feduc.2024.1403745
  3. European Commission. (2022). Ethical guidelines on the use of artificial intelligence (AI) and data in teaching and learning for educators. Publications Office of the European Union. https://doi.org/10.2766/153756
  4. Li, D. (2026). Research on the application of AI pitch correction system in children’s music teaching. Frontiers in Artificial Intelligence and Applications. https://doi.org/10.3233/faia260353
  5. 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
  6. Miao, F., y Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO. https://doi.org/10.54675/EWZM9535
  7. 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
  8. 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
  9. OECD. (2021). Starting Strong VI: Supporting meaningful interactions in early childhood education and care. OECD Publishing. https://doi.org/10.1787/f47a06ae-en
  10. OECD. (2023). Empowering young children in the digital age (Starting Strong). OECD Publishing. https://doi.org/10.1787/50967622-en
  11. Roldan-Cardona, M., Chacón-Castro, M., Jadán-Guerrero, J., Salvador-Ullauri, L., y Acosta-Vargas, P. (2025). Music education with artificial intelligence for inclusive and sustainable early childhood learning. Emerging Science Journal, 9, 201–218. https://doi.org/10.28991/ESJ-2025-SIED1-012
  12. Şenol, F. B., y Karaca, N. H. (2025). Investigation of the effect of music education programme on preschoolers’ motor creativity skills. Education 3-13, 53(2), 252–265. https://doi.org/10.1080/03004279.2023.2172356
  13. Shi, X. (2026). An affective computing framework for preschool music education using MoE architecture and neural radiance fields. Discover Artificial Intelligence, 6, 812. https://doi.org/10.1007/s44163-026-01858-0
  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. UNESCO. (2021). Recommendation on the ethics of artificial intelligence. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000381137
  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. Webster, K., y Holmes, A. (2026). Reconceptualizing music as a key pedagogical tool in early childhood education: An early years case study. Journal of Research in Childhood Education. https://doi.org/10.1080/02568543.2026.2664892