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 dialogic reading / shared book reading: a chatbot/LLM agent that “does dialogic reading” with the child in place of the adult (StoryPal, Storio/Mia or another agent that asks CROWD/PEER questions without human co-presence); a voice/AI system that scores “reading comprehension” / “vocabulary growth score” / “dialogic quality score” / “reading engagement score”; a GenAI that generates “CROWD/PEER scripts” or “dialogic reading lesson plans” by prompt as a closed recipe the adult reads aloud without contingency; a dashboard of “dialogic reading minutes” / “question quality score” / “shared reading minutes” for administrative surveillance; and an app that replaces adult–child shared reading with a conversational agent. The five allow centers to display that they “already do dialogic reading with AI.” The leap —from agent that questions the child, comprehension/vocabulary/dialogic-quality score, GenAI script, minutes metric or substitute app to claiming support— is not authorized by AI-in-ECE mappings nor by the pedagogy of dialogic reading when there is a book (physical or e-book) in adult co-presence that makes contingent PEER/CROWD moves, listens, evaluates and expands, and child agency to answer and ask.

The thesis of this article is restrictive. A chatbot/LLM agent that “does dialogic reading” with the child in place of the adult, a voice/AI system that scores “reading comprehension” / “vocabulary growth score” / “dialogic quality score” / “reading engagement score” for the 3–6-year-old, a GenAI that generates “CROWD/PEER scripts” or “dialogic reading lesson plans” by prompt as a closed recipe without contingency, a dashboard of “dialogic reading minutes” / “question quality score” / “shared reading minutes” for administrative surveillance, or an app that replaces adult–child shared reading with a conversational agent do not constitute support for dialogic reading in the early years. In early childhood dialogic reading develops with a book (physical or e-book) in adult co-presence that makes contingent PEER/CROWD moves, listens, evaluates and expands (Kennedy & McLoughlin, 2022; Dicataldo, Rowe & Roch, 2022; Yang et al., 2022); AI may support adult preparation, not replace the relationship nor turn the child into a score object. Kennedy and McLoughlin (2022) systematically review dialogic reading with English learners and anchor PEER/CROWD; ELL/EFL transfer is marked without converting this article into multilingualism. Dicataldo et al. (2022) document a parent-focused intervention: craft runs through the adult who does DR. Xiao, Zou, Lin, Li and Yang (2025) provide an RCT of parent-led versus AI-guided dialogic reading in a children’s e-book context: marked central empirical finding with limits; it does not authorize replacing the adult in a 3–6 classroom. Xiao, Li, Lin, Zou, Yang, Zou and Xiong (2025) describe Storio/Mia with N = 17 EFL children plus parents and educators: sample limit and EFL transfer; the agent does not sign adult co-presence in CENDI. Zhang, Albashtawi and Mahfoodh (2026) —N = 108; Chinese bilingual home— mark AI-assisted versus traditional DR: “assisted” ≠ replacement. That authorizes asking what was measured: agent turns, score, script by prompt, minutes or substitute app —not the practice when an adult and a child converse about text and images with contingent PEER/CROWD.

The problem is worsened by six category confusions. First: dialogic reading is not generic literacy —there may be emergent literacy IN DR (Kennedy & McLoughlin, 2022); the axis is the adult–child relational practice around the book—. Second: it is not generic orality —there is talk, but the axis is dialogic turns about the book, not “talking more”—. Third: it is not multilingualism as axis —EFL/bilingual sources (Xiao et al., 2025 CAEAI; Zhang et al., 2026; Yang et al., 2022; Kennedy & McLoughlin, 2022) are marked as transfer; this article is not converted into multilingualism already published—. Fourth: it is not generic parental mediation as axis —Dicataldo et al. (2022) and the parent-led arm of Xiao et al. (2025, BJET) anchor the adult in DR; parental mediation already published is not recycled as the theme—. Fifth: it is not AI tutors or reading-interest chatbots —Liu et al. (2022) are read as chatbot–interest contrast, not adult–child DR—. Sixth: it is not LLM story personalization as HCI —Chen et al. (2025, CHI) is a peripheral multi-stakeholder contrast, not DR craft in ECE. This work does not recycle literacy, orality, student alphabetization, multilingualism, parental mediation as axis, AI tutors, symbolic play, joint attention, adult–child interactions as generic axis, documentation, assessment, SEL, teacher education, sand/water/playdough/loose parts/outdoor/risky/block/spatial, STEM, music, numeracy, UDL, low connectivity or privacy. Galbraith (2022) is a peripheral contrast of play pedagogies; it is not converted into teacher education already published. The question is what counts as support for dialogic reading when a center “does AI and dialogic reading.”

There is, moreover, an economy of coordination: the five artifacts fit on a slide; the prolonged episode with book, contingent PEER/CROWD turns, adult listening that evaluates and expands, and child agency to answer and ask, does not. NAEYC (2022) and OECD (2021, 2023) require meaningful interactions and subordinated digitalization. Inference: support is not fulfilled when an agent asks CROWD of the child without co-presence, when vocabulary growth or dialogic quality is scored, or when scripts are generated by prompt without contingency. The contributions are three: separate craft from artifacts; examine marked cases (Kennedy & McLoughlin, 2022; Dicataldo et al., 2022; Xiao et al., 2025 BJET; Xiao et al., 2025 CAEAI; Zhang et al., 2026; Yang et al., 2022; Liu et al., 2022 as contrast; Chen et al., 2025 and Galbraith, 2022 as peripheral); and offer four tests of support for dialogic reading.

2. State of the art: from situated dialogic-reading practice to the artifact on display

It is useful to separate four strata that the “AI for dialogic reading in ECE” market often blends. The first is the construct of dialogic reading / shared book reading ages 3–6 as adult–child relational practice around a book (physical or e-book): PEER and CROWD turns, co-presence that listens and expands, child agency (Kennedy & McLoughlin, 2022; Dicataldo et al., 2022; Yang et al., 2022). The second is the craft that cultivates it —shared book, adult who Prompt–Evaluate–Expand–Repeat contingently, CROWD according to the episode, not substituting the relationship with an agent nor turning the child into a score object— (Kennedy & McLoughlin, 2022; Dicataldo et al., 2022; Xiao et al., 2025 BJET, parent-led; NAEYC, 2022; OECD, 2021, 2023). The third is evidence on AI affordances in ECE and LLM reading systems —without equating them to situated DR— (Chen, 2024; Su & Yang, 2022; Su & Zhong, 2022; Ljungcrantz, 2026; Nikolopoulou, 2025; Xiao et al., 2025 CAEAI; Zhang et al., 2026; Liu et al., 2022; Chen et al., 2025). The fourth is the rights and DAP framework that treats the child as subject of conversation about the book, not as a vector of vocabulary growth score nor exclusive recipient of an agent without an adult (UNESCO, 2021; Miao & Holmes, 2023; U.S. Department of Education, 2023).

In construct and craft, Kennedy and McLoughlin (2022) saturate the floor: systematic review of dialogic reading with ELL; anchors PEER/CROWD and adult-mediated shared reading; ELL transfer. Dicataldo et al. (2022) provide a parent-focused intervention: craft is taught to the adult. Yang et al. (2022) examine bilingual discussion prompts in shared e-book reading: prompts that mediate conversation, not an autonomous score; bilingual transfer. Xiao et al. (2025, BJET) contrast parent-led and AI-guided: RCT; the parent-led arm preserves co-presence; AI-guided alone does not sign DR support in a 3–6 classroom without an adult. Xiao et al. (2025, CAEAI) —N = 17 EFL; Storio/Mia—: sample and EFL limits; do not equate to adult replacement. Zhang et al. (2026) —N = 108; Chinese bilingual home— mark “assisted” ≠ replacement. Inference: a dialogic quality score does not listen or expand; an adult who does contingent PEER does.

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 DR. Nikolopoulou (2025) balances promises and challenges of child-centered GenAI under teacher mediation: caution framework. Liu et al. (2022) analyze children’s interaction with a chatbot and reading interest: empirical contrast —interest ≠ adult–child DR—. Chen et al. (2025) characterize LLM-empowered personalized story reading from multi-stakeholder perspectives at CHI: peripheral HCI contrast, not DR craft in CENDI. 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 “CROWD/PEER scripts” or “dialogic reading lesson plans” 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. Galbraith (2022) offers a peripheral contrast of play pedagogies among preservice teachers: not converted into a teacher-education axis.

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 classroom ages 3–6 or is EFL/home/bilingual; (b) dialogic reading, shared book reading, PEER/CROWD, shared e-book, or AI in ECE with artifact/craft relevance; (c) kindergarten, preschool, CENDI or ages 3–6, or explicit transfer; (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 —generic literacy, orality, student alphabetization, multilingualism, parental mediation as axis, AI tutors, symbolic play, joint attention, adult–child interactions as generic axis, documentation, assessment, SEL, teacher education, sand/water/playdough/loose parts/outdoor/risky/block/spatial, STEM, music, numeracy, UDL, low connectivity and privacy—. Galbraith (2022) only as peripheral contrast of play pedagogies. Chen et al. (2025) = peripheral HCI contrast. Liu et al. (2022) = chatbot–interest contrast, not DR. EFL/bilingual/home sources (Kennedy & McLoughlin, 2022; Xiao et al., 2025 CAEAI; Zhang et al., 2026; Yang et al., 2022) with marked transfer; they do not convert the axis into multilingualism.

The search was executed on 5 September 2026 (slot 13:02 America/Mexico_City) on DOI pages, Crossref, Springer, Elsevier, Wiley, Taylor & Francis, MDPI, ACM DL, 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 comprehension/vocabulary growth/dialogic quality/reading engagement scoring metrics, shared-reading-minutes dashboard or total-replacement app in CENDI classroom 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 agent/chatbot empirical contrasts (Xiao et al., 2025 CAEAI; Liu et al., 2022; Xiao et al., 2025 BJET). The twenty verified corpus sources were used.

4. Case 1. Chatbot/LLM agent that does DR with the child, comprehension/vocabulary/dialogic quality/engagement score, GenAI of CROWD/PEER scripts or lesson plans, minutes/question-quality dashboard, or app that replaces shared reading do not constitute support for dialogic reading

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 dialogic reading. Chen (2024) maps global AI affordances in early childhood education: scoping finding on emerging uses —tutoring, analytics, content generation, agents—, not a finding that an LLM agent that asks CROWD of the child without an adult cultivates situated DR. Su and Yang (2022) review the AI-in-ECE field: synthesis finding on trends, not on co-presence that listens, evaluates and expands in PEER turns. Ljungcrantz (2026) reviews AI–ECE interaction 2020–2024: state-of-the-art finding, not mediated shared book reading. Pedagogical inference, marked as such: the gesture “the child had a high dialogic quality score = there was support for dialogic reading” is a conversational-quality analytics ceiling. Kennedy and McLoughlin (2022) and Dicataldo et al. (2022) ask for an adult who does contingent DR; a score useful for administration can coexist with absence of co-presence and of real child agency.

Nikolopoulou (2025) and Miao and Holmes (2023) name the risk of the generator of “CROWD/PEER scripts” or “dialogic reading lesson plans” from a prompt that closes contingency. 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 DR episode with book and adult who listens and expands (Kennedy & McLoughlin, 2022; Dicataldo et al., 2022). Inference: producing a CROWD/PEER script by prompt may be subordinated teacher preparation; support for dialogic reading begins when there is a shared book, an adult who Prompt–Evaluate–Expand–Repeat contingently to the child’s response, and child agency to ask as well. Su and Zhong (2022) propose AI curriculum design in ECE as a future direction —AI literacy curriculum, not situated DR—. Limit inference: an AI curriculum does not sign conversation about the book.

The chatbot/LLM agent that “does DR” with the child (Storio/Mia or analogues), the voice/AI system that scores comprehension/vocabulary/dialogic quality/engagement, the minutes/question-quality dashboard and the app that replaces shared reading lack, in the verified corpus, trials with N, d, r or AUC invented here to declare that the score or the agent constitute the craft in CENDI ages 3–6. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map analytics, tutoring and agents without equivalence to mediated DR. Xiao et al. (2025, CAEAI) —N = 17 EFL; Storio/Mia—: marked finding; do not authorize replacing co-presence. Xiao et al. (2025, BJET): RCT; AI-guided does not erase the parent-led arm nor sign classroom replacement. Zhang et al. (2026) —N = 108— mark home assistance, not minutes surveillance in kindergarten. Liu et al. (2022): chatbot–interest contrast ≠ PEER/CROWD with adult. Chen et al. (2025): peripheral HCI contrast. Restrictive inference: agent ≠ adult who listens and expands; GenAI scripts ≠ contingent Kennedy/Dicataldo; vocabulary growth score ≠ co-presence; dashboard ≠ observation of the turn; replacement app ≠ shared book. UNESCO (2021) and U.S. Department of Education (2023) require human oversight. The five artifacts share one substitution grammar: the agent speaks for the adult; the score speaks for the child; GenAI closes contingency; the dashboard substitutes observation; the app replaces the relationship.

It is useful to specify the ceiling without inventing effects. An agent can ask Wh-questions; Kennedy and McLoughlin (2022) articulate contingent human CROWD, not AUC. A voice system can score engagement; Dicataldo et al. (2022) teach the adult to do DR. A GenAI can print PEER scripts; Miao and Holmes (2023) require validation. A dashboard can count minutes; NAEYC (2022) and OECD (2021) anchor practice. An app can converse with the child; Xiao et al. (2025, BJET) and OECD (2023) do not authorize erasing the adult. Inference: “high dialogic quality” or “lesson plan according to the model” does not sign support.

5. Case 2. What the kindergarten does when there is support for dialogic reading: PEER/CROWD, Kennedy and McLoughlin, Dicataldo et al., Xiao et al. (BJET and CAEAI), Zhang et al., Yang et al. —with Liu et al., Chen et al. (CHI) and Galbraith as contrasts

Kennedy and McLoughlin (2022) saturate the craft floor: systematic review of dialogic reading to develop emergent literacy in ELL; they anchor PEER and CROWD as adult moves in shared reading, not as dashboard metrics. Status: systematic-review framework; marked ELL transfer; the article is not converted into multilingualism. Pedagogical inference: a center may call it support for dialogic reading when it protects a shared book, contingent adult co-presence and child agency —not when an agent questions the child alone—. Dialogic reading is not a vocabulary growth score: it is Prompt, Evaluate, Expand, Repeat; Completion, Recall, Open-ended, Wh-questions, Distancing; listening to the child’s response and expanding without a closed recipe.

Dicataldo, Rowe and Roch (2022) provide marked parental craft: “Let’s read together” intervention on dialogic book reading in preschool. Status: empirical finding; the object is the adult who does DR, not an agent. Inference: craft is cultivated by teaching the adult; it is not signed by GenAI scripts without contingency. Xiao, Zou, Lin, Li and Yang (2025) provide the RCT anchor of parent-led vs AI-guided in e-book (BJET). Limit: e-book; d and r are not invented here; it does not authorize replacing the adult in CENDI ages 3–6. The parent-led arm preserves the category; AI-guided is read as subordinable guidance, not replacement. Xiao, Li, Lin, Zou, Yang, Zou and Xiong (2025) —CAEAI; Storio/Mia; N = 17 EFL plus parents and educators—: small sample; EFL transfer; the agent does not replace co-presence. Zhang, Albashtawi and Mahfoodh (2026) —N = 108; AI-DR in Chinese bilingual home— mark “assisted” ≠ replacement nor minutes dashboard in kindergarten.

Yang, Xia, Collins and Warschauer (2022) examine bilingual discussion prompts in shared e-book reading: prompts that mediate conversation, not an autonomous score; bilingual transfer without converting the axis into multilingualism. Liu, Liao, Chang and Lin (2022): chatbot and reading interest —contrast; interest ≠ DR with PEER/CROWD—. Chen et al. (2025, CHI): multi-stakeholder LLM personalization —peripheral HCI contrast—. Galbraith (2022) —play pedagogies among preservice teachers— is a peripheral contrast; not teacher education as axis. NAEYC (2022) and OECD (2021, 2023) anchor meaningful interactions and subordinated digitalization. Mediation coherent with DR —Kennedy & McLoughlin (2022); Dicataldo et al. (2022); parent-led of Xiao et al. (2025, BJET); Yang et al. (2022)— is co-presence that offers a book, makes contingent PEER/CROWD moves, listens, evaluates, expands and observes child agency without a closed digital script. Inference: pedagogical observation reads the turn; the dashboard counts minutes. A dashboard does not Expand; an adult who listens does.

6. Case 3. Dialogic reading ≠ generic literacy, orality, multilingualism as axis, generic parental mediation, AI tutors or reading-interest chatbots

The first frontier is generic literacy: Kennedy and McLoughlin (2022) show emergent literacy IN dialogic reading; they do not authorize converting this article into the literacy axis already published nor a reading engagement score as if it were literacy. Inference: “we work on literacy = there is dialogic reading” does not sign. A vocabulary growth score is analytics, not contingent PEER. The second is orality: there is speech in CROWD and Expand; the axis is turns about the book, not generic orality already published. The third is multilingualism as axis: EFL/bilingual sources (Kennedy & McLoughlin, 2022; Xiao et al., 2025 CAEAI; Zhang et al., 2026; Yang et al., 2022) are marked as transfer; they do not authorize rewriting the article as multilingualism. The fourth is generic parental mediation: Dicataldo et al. (2022) and parent-led of Xiao et al. (2025, BJET) anchor the adult in DR; parental mediation already published is not recycled as the central theme. The fifth is AI tutors: Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map tutoring as an affordance; here the axis is DR, not a generic tutor. The sixth is reading-interest chatbots: Liu et al. (2022) contrast interest; Chen et al. (2025) contrast HCI personalization; neither signs adult–child shared book reading. UNESCO (2021), Miao and Holmes (2023) and U.S. Department of Education (2023) require human oversight. Inference: coherent AI use at ages 3–6 remains on the adult’s side —preparation, subordinated prompts (Yang et al., 2022), Galbraith (2022) contrast of situated pedagogy— contrasted with Kennedy/Dicataldo and the parent-led arm. It does not enter as an agent that does DR with the child, dialogic-quality scorer, GenAI of closed scripts, minutes dashboard or replacement app.

7. Inferential framework: four tests to affirm that there is support for dialogic reading, 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 dialogic reading.

7.1. Test of situated practice of shared book (physical or e-book), contingent PEER/CROWD turns and child agency, not of the LLM agent that “does DR” with the child nor of the app that replaces shared reading. Kennedy and McLoughlin (2022), Dicataldo et al. (2022), Yang et al. (2022) and parent-led of Xiao et al. (2025, BJET) define craft as adult–child conversation about the book. Chen (2024), Su and Yang (2022), Ljungcrantz (2026), Xiao et al. (2025, CAEAI) and Liu et al. (2022) map agents/chatbots as affordance or contrast, not as complete situated DR. Inference: support is verified in shared book, adult who does contingent PEER and child with agency. If the center’s “evidence” is an agent that asks CROWD alone or an app that replaces the adult, there is conversational substitution, not support for dialogic reading.

7.2. Test of adult co-presence that listens, evaluates and expands, not of GenAI “CROWD/PEER scripts” or “dialogic reading lesson plans” by prompt. Kennedy and McLoughlin (2022), Dicataldo et al. (2022) and NAEYC (2022) situate contingency, expand and developmentally appropriate practice. Miao and Holmes (2023) require pedagogical validation of generative AI. Inference: producing scripts or lesson plans by prompt that close contingency does not demonstrate that an adult listened to the child’s response and expanded. A script may exist; it does not sign the DR episode.

7.3. Test of human mediation —including marked AI-assisted (Zhang et al., 2026) and parent-led (Xiao et al., 2025 BJET)—, not of the minutes/question-quality dashboard nor of the comprehension/vocabulary/dialogic quality/engagement score. OECD (2021, 2023) require meaningful interactions and subordinated digitalization. Zhang et al. (2026) anchor home assistance with transfer, not minutes surveillance in kindergarten. Xiao et al. (2025, BJET) anchor the adult in parent-led. Inference: “high minutes” or “high dialogic quality score” may raise an administrative threshold without raising co-presence or agency. Pedagogical observation reads the turn; the dashboard counts surveillance. Adult preparation (Dicataldo et al., 2022; Yang et al., 2022; Galbraith, 2022 peripheral) is legitimate; the scorer that turns the child into a metric is not.

7.4. Test of category distinction and professional judgment, not of the product catalogue. Dialogic reading ≠ generic literacy, orality, multilingualism as axis, generic parental mediation, AI tutors or reading-interest chatbots. Kennedy and McLoughlin (2022) prevent reducing DR to literacy already published. Liu et al. (2022) and Chen et al. (2025) prevent confusing reading interest or HCI personalization with adult–child shared book reading. 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 vector of vocabulary growth score nor as exclusive recipient of an agent that “does DR.” Support for dialogic reading is not fulfilled by scoring “dialogic quality” better or surveilling “minutes.” It is fulfilled by practicing shared book, adult co-presence that makes contingent PEER/CROWD moves, listens, evaluates and expands, and child agency to answer and ask.

The framework admits digital subordinated to adult preparation and validated human mediation (Dicataldo et al., 2022; Yang et al., 2022; Zhang et al., 2026 as marked AI-assisted; Chen, 2024; Su & Yang, 2022; Ljungcrantz, 2026). It rejects declaring support via the five artifacts (Miao & Holmes, 2023; Nikolopoulou, 2025; UNESCO, 2021; Xiao et al., 2025 BJET and CAEAI with limits). 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 dialogic reading. Kennedy and McLoughlin (2022), Dicataldo et al. (2022), Xiao et al. (2025, BJET and CAEAI; N = 17 EFL), Zhang et al. (2026; N = 108), Yang et al. (2022), Liu et al. (2022), Chen et al. (2025) and Galbraith (2022) sustain the craft or its boundary; Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) map AI without equivalence to situated DR.

The second is between automated assessment of comprehension/vocabulary/dialogic quality/engagement or minutes and pedagogy of the dialogic turn. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) objectivize analytics; AUC of dialogic quality scoring in CENDI ages 3–6 is not invented here. Inference: declaring support via vocabulary growth or question quality score is inverted pedagogy.

The third is between GenAI scripts / agent / replacement app and integral craft. Miao and Holmes (2023) and Nikolopoulou (2025) set limits; Kennedy and McLoughlin (2022), Dicataldo et al. (2022) and parent-led of Xiao et al. (2025, BJET) set adult contingency. Inference: selling scripts by prompt, Storio/Mia as substitute or a replacement app as “dialogic reading with AI” confuses product with relational practice. Zhang et al. (2026) confirm AI-assisted N = 108 with transfer without authorizing replacement; UNESCO (2021), Miao and Holmes (2023) and U.S. Department of Education (2023) subordinate AI to professional judgment.

The four tests of section 7 read these tensions. The empirical contrast 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 (Dicataldo et al., 2022; Yang et al., 2022; Zhang et al., 2026); inverting the sequence is not (NAEYC, 2022; OECD, 2021, 2023; Kennedy & McLoughlin, 2022).

9. Limits

This review is narrative. It does not apply its own PRISMA nor estimate combined effects. Xiao et al. (2025, CAEAI) measure N = 17 EFL: limit and marked transfer. Zhang et al. (2026) measure N = 108 in a Chinese bilingual home: marked transfer. Xiao et al. (2025, BJET) provide an RCT in e-book parent-led vs AI-guided: context limits; d and r are not invented here. Kennedy and McLoughlin (2022) provide an ELL systematic review: transfer. Dicataldo et al. (2022) and Yang et al. (2022) delimit craft and intervention/prompts. Liu et al. (2022), Chen et al. (2025) and Galbraith (2022) are contrasts (chatbot–interest; HCI; play pedagogies), not axes. Chen (2024), Su and Yang (2022), Su and Zhong (2022), Ljungcrantz (2026) and Nikolopoulou (2025) map AI in ECE, not dialogic quality scores in CENDI. Verified trials of the five artifacts as a display package in Latin American CENDI ages 3–6 were not 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 chatbot/LLM agent that “does dialogic reading” with the child in place of the adult, a voice/AI system that scores “reading comprehension” / “vocabulary growth score” / “dialogic quality score” / “reading engagement score”, a GenAI that generates “CROWD/PEER scripts” or “dialogic reading lesson plans” by prompt without contingency, a dashboard of “dialogic reading minutes” / “question quality score” / “shared reading minutes”, or an app that replaces shared reading do not constitute support for dialogic reading in early childhood education. Chen (2024), Su and Yang (2022) and Ljungcrantz (2026) confirm AI affordances without equivalence to situated DR. Nikolopoulou (2025) and Miao and Holmes (2023) set GenAI limits. Xiao et al. (2025, CAEAI; N = 17 EFL) and Xiao et al. (2025, BJET) require reading agents and AI-guided with limits. When there is support, there is situated practice: PEER/CROWD (Kennedy & McLoughlin, 2022); intervention with the adult (Dicataldo et al., 2022); parent-led (Xiao et al., 2025, BJET); mediating prompts (Yang et al., 2022); marked AI-assisted (Zhang et al., 2026); contrasts Liu et al. (2022), Chen et al. (2025) and Galbraith (2022); DAP and interactions (NAEYC, 2022; OECD, 2021, 2023). Dialogic reading is distinguished from generic literacy, orality, multilingualism as axis, generic parental mediation, AI tutors and interest chatbots. 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 claim it. Where they measure AI mappings, GenAI, EFL agents, home AI-DR or interest chatbots, it does not translate them into pedagogical dialogic-reading support via score or substitute agent. Accompanying children aged three to six in dialogic reading is to exercise a shared book (physical or e-book), adult co-presence that makes contingent PEER/CROWD moves, listens, evaluates and expands, and child agency to answer and ask. The rest is a chatbot/LLM agent that does DR with the child, a comprehension/vocabulary/dialogic quality/engagement score, GenAI of CROWD/PEER scripts, a minutes dashboard and an app that replaces shared reading. It is not support for dialogic reading 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., Tang, M., Lu, Y., Yao, B., Fan, E., Ma, X., Xu, Y., Wang, D., Sun, Y., y He, L. (2025). Characterizing LLM-empowered personalized story reading and interaction for children: Insights from multi-stakeholder perspectives. En Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (pp. 1–24). ACM. https://doi.org/10.1145/3706598.3713275
  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. Dicataldo, R., Rowe, M. L., y Roch, M. (2022). “Let’s read together”: A parent-focused intervention on dialogic book reading to improve early language and literacy skills in preschool children. Children, 9(8), Article 1149. https://doi.org/10.3390/children9081149
  4. 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
  5. Kennedy, C., y McLoughlin, A. (2022). Developing the emergent literacy skills of English language learners through dialogic reading: A systematic review. Early Childhood Education Journal, 51(2), 317–332. https://doi.org/10.1007/s10643-021-01291-1
  6. Liu, C.-C., Liao, M.-G., Chang, C.-H., y Lin, H.-M. (2022). An analysis of children’ interaction with an AI chatbot and its impact on their interest in reading. Computers & Education, 189, Article 104576. https://doi.org/10.1016/j.compedu.2022.104576
  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. Su, J., y Yang, W. (2022). Artificial intelligence in early childhood education: A scoping review. Computers and Education: Artificial Intelligence, 3, Article 100049. https://doi.org/10.1016/j.caeai.2022.100049
  14. Su, J., y Zhong, Y. (2022). Artificial Intelligence (AI) in early childhood education: Curriculum design and future directions. Computers and Education: Artificial Intelligence, 3, Article 100072. https://doi.org/10.1016/j.caeai.2022.100072
  15. 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
  16. UNESCO. (2021). Recommendation on the ethics of artificial intelligence. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000381137
  17. Xiao, F., Li, Z., Lin, J., Zou, X., Yang, D., Zou, W., y Xiong, J. (2025). Leveraging an LLM-enhanced bilingual conversational agent for EFL children’s dialogic reading: Insights from children, parents, and educators. Computers and Education: Artificial Intelligence, 9, Article 100484. https://doi.org/10.1016/j.caeai.2025.100484
  18. Xiao, F., Zou, E. W., Lin, J., Li, Z., y Yang, D. (2025). Parent-led vs. AI-guided dialogic reading: Evidence from a randomized controlled trial in children’s e-book context. British Journal of Educational Technology, 56(5), 1784–1813. https://doi.org/10.1111/bjet.13615
  19. Yang, D., Xia, C., Collins, P., y Warschauer, M. (2022). The role of bilingual discussion prompts in shared e-book reading. Computers & Education, 190, Article 104622. https://doi.org/10.1016/j.compedu.2022.104622
  20. Zhang, D., Albashtawi, A. H., y Mahfoodh, O. H. A. (2026). AI-assisted dialogic reading at home: Enhancing bilingual expressive vocabulary and engagement among Chinese pre-schoolers. Early Child Development and Care, 196(5–6), 424–445. https://doi.org/10.1080/03004430.2026.2667895