Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 64 of the 64 references it cites were found in a registry or an index; 11 of the 35 sources its claims cite were checked for support.

Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review

Open access

Abstract

Objective: Despite the high demand for speech and language therapy (SLT) for children with speech sound disorders (SSDs), accessible services remain limited. Technology-driven efforts have led to the development of systems and applications to assist children, parents, and therapists in the SLT process. AI and machine learning (ML), particularly through automatic speech recognition and audio processing techniques, play a central role in these advancements. This scoping review examines studies focusing on these techniques for managing the SLT process. Methods: To include the most relevant studies, a systematic search was conducted on 3 February 2025 across five major databases (PubMed, Scopus, ScienceDirect, ACM Digital Library, and IEEE Xplore), following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Scoping Reviews guidelines. After applying our criteria, 30 of the 188 identified studies met the eligibility requirements. Results: These studies predominantly utilize deep neural networks, ML classifiers, acoustic features, and audio processing techniques to detect SSDs. The findings demonstrate the effectiveness of these applications to support therapists in diagnostics. Moreover, computer-based tools have proven more engaging for children than traditional therapy by offering personalized therapy plans and real-time feedback. These systems enable therapists to monitor progress and adjust treatments. Conclusion: This review provides an overview of AI-assisted SLT models, highlights gaps, and suggests directions for future research. It shows the effectiveness and potential of AI in enhancing the SLT process. However, challenges related to data privacy, accessibility, and the need for clinical validation persist and need to be addressed in the future.

Overview

CoChat Summary

## TL;DR

This scoping review maps AI-driven technologies for pediatric speech and language therapy (SLT), synthesizing 30 eligible studies from 188 identified records across five databases. It finds promising AI support for SSD detection, ASR adaptation, and child engagement, but highlights major gaps in data scarcity, multilingual support, clinical validation, and severity/type classification.

## Key findings

  • The search on 3 February 2025 across PubMed, Scopus, ScienceDirect, ACM Digital Library, and IEEE Xplore identified 188 papers; after removing 64 duplicates and applying eligibility criteria, 30 studies were included.
  • The included studies mainly fell into two technical goals: SSD detection and SLT games; most studies focused on disorder detection, and most systems used binary output rather than classifying disorder type or severity.
  • The review reports that the most common approaches were traditional ML classifiers (especially SVM) and deep learning/ASR methods; among the reviewed papers, SVM was the most frequently used classifier in papers A, C, G, I, K, M, P, R, S, U, and Y.
  • Participant ages ranged from 2 to 16 years, with an average focus on children aged 4 to 10 years; reported sample sizes ranged from 12 in paper W to 1362 in paper U, and the papers J, O, P, Q, R, S, and Y reported 231 boys (63%) and 137 girls (37%).

## Methods

  • Scoping review following PRISMA-ScR guidelines, with a systematic search conducted on 3 February 2025 in five databases: PubMed, Scopus, ScienceDirect, ACM Digital Library, and IEEE Xplore.
  • Included studies were charted in a spreadsheet by author, year, country, study aims, methods, results, and conclusions; 30 studies met criteria after full-text screening.
  • The review addressed six research questions covering technical goals, languages and participant selection, pathologies, application features, AI models/speech properties, and datasets.

## Limitations

  • The paper states that it was restricted to English-language articles, which may have missed relevant work published in other languages.
  • The paper states that it focused mainly on papers published after 2018, and its keyword strategy may have excluded relevant studies that did not explicitly mention “children” in the title.
  • Evident from the text: the review is limited to AI-based systems and does not evaluate pathological components or whether therapy methods are well aligned with the technology, which the authors explicitly note lies outside their expertise.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1177/20552076251376533

Cites

64 of 64 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram illustrating the paper selection process.
Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram illustrating the paper selection process.CC BY · PubMed Central
Figure 2. The publication years of the included studies. Stage 1 represents the initial count of papers identified by our search terms, following the removal of duplicates. Stage 2 indicates the number of papers retained after an initial screening, which excluded those that were not relevant. Stage 3 reflects the final count of studies included after a comprehensive review and application of the exclusion criteria.
Figure 2. The publication years of the included studies. Stage 1 represents the initial count of papers identified by our search terms, following the removal of duplicates. Stage 2 indicates the number of papers retained after an initial screening, which excluded those that were not relevant. Stage 3 reflects the final count of studies included after a comprehensive review and application of the exclusion criteria.CC BY · PubMed Central
Figure 3. High-level overview of the reviewed approaches, illustrating the key components and their interrelationships.
Figure 3. High-level overview of the reviewed approaches, illustrating the key components and their interrelationships.CC BY · PubMed Central

Verification

Claims the paper makes

9 Valid · 7 Not valid · 22 Not checked · 111 Own results and interpretations

40 more claims

Create a free account to see all 149 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 74
  • It may state two separate claims as one. 58
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Citations

Cites

Show all 59 references

Research works printed but not linked: 4

  1. Radford A Kim JW Xu T et al. Robust speech recognition via large-scale weak supervision. In: Proceedings of the 40th International conference on machine learning (ICML) Honolulu HI USA; Proceedings of Machine Learning Research (PMLR) Scotts Valley CA 23–29 July 2023 pp.28492–28518. PMLR.
  2. Bhat HF. Evaluating SVM algorithms for bioinformatics gene expression analysis. Int J Comput Sci Eng 2017; 6: 42–52.
  3. Fohr D Mella O Illina I. New paradigm in speech recognition: Deep neural networks. In: Seventh international conference on information systems and economic intelligence (SIIE) Al-Hoceima Morocco 2017 May 11–13 pp.1–6. Piscataway NJ: IEEE.
  4. Howitt AW. Automatic syllable detection for vowel landmarks. ScD Thesis Massachusetts Institute of Technology Cambridge MA USA 2000.

Cited by

See all 4

Metadata

Authors
Dadgar, M., Ennis, C., Mokgosi, K., & Ross, R.
Affiliations
  • Technological University Dublin (IE)
  • National University of Ireland, Maynooth (IE)
Corresponding author
Dadgar, Milad
Venue
DIGITAL HEALTH
Published
2025-05
Volume
11
Pages
20552076251376533-20552076251376533
ISSN
2055-2076 (ISSN-L)
OpenAlex
W4416168658
Cited by
4
References
59
Citations by year
2026: 4
FWCI
0.99
Type
journal article
Language
en
Open access
Licence
Locations
  • Digital Health (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Maynooth University ePrints and eTheses Archive (Maynooth University) (acceptedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Science Foundation Ireland (18/CRT/6183)
Topics
  • Social Sciences › Psychology › Developmental and Educational Psychology › Language Development and Disorders
  • Health Sciences › Health Professions › Speech and Hearing › Voice and Speech Disorders
  • Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
Keywords
artificial intelligence, speech therapy, speech recognition, machine learning, deep neural network, acoustic features, computer-based therapy, real-time feedback, data privacy

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review. https://app.cochat.ai/papers/10.1177/20552076251376533
The paper
Dadgar, M., Ennis, C., Mokgosi, K., & Ross, R. (2025). Artificial intelligence (AI)-driven technologies for managing pediatric speech and language therapy: A scoping review. DIGITAL HEALTH. https://doi.org/10.1177/20552076251376533

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.