ChatGPT As a Resource to Strengthen Mathematics Teaching and Learning: A Systematic Review.
Background: Since late 2022, the integration of generative AI, particularly ChatGPT, has rapidly expanded in mathematics education, creating a dynamic but heterogeneous research field that requires systematic synthesis regarding its pedagogical value. Objective: This review analyses scientific liter...
| Publicado en: | Journal of Computer Assisted Learning Vol. 42; no. 3; pp. 1 - 27 |
|---|---|
| Autores principales: | , , |
| Formato: | research systematic review tables/charts Journal Article |
| Publicado: |
Wiley-Blackwell
Jun2026
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=194050885&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 194050885 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 02664909 6M1 jtl: Journal of Computer Assisted Learning issn: 02664909 maglogo: Y pubinfo: dt: Jun2026 vid: 42 iid: 3 pid: 480 pub: Wiley-Blackwell place: Malden, Massachusetts artinfo: ui: 194050885 194050885 194050885 10.1002/jcal.70241 194050885 ppf: 1 ppct: 26 formats: tig: atl: ChatGPT As a Resource to Strengthen Mathematics Teaching and Learning: A Systematic Review. aug: au: Su, Chia Shih Hsu, Chuan Chih González Campos, José Alejandro affil: Facultad de Ciencias Básicas, Universidad Católica del Maule (UCM), Talca, Chile sug: subj: Artificial Intelligence, Generative Mathematics Education Teaching Learning Human Systematic Review Problem Solving Cognition Descriptive Statistics Data Analysis Software Thematic Analysis Inferential Statistics Content Analysis ab: Background: Since late 2022, the integration of generative AI, particularly ChatGPT, has rapidly expanded in mathematics education, creating a dynamic but heterogeneous research field that requires systematic synthesis regarding its pedagogical value. Objective: This review analyses scientific literature on ChatGPT in mathematics education to identify its applications, benefits, limitations and instructional projections. Methods: Following PRISMA guidelines, 54 peer‐reviewed articles from Web of Science, Scopus and SciELO were analysed using deductive‐inductive content analysis across methodological, pedagogical and impact dimensions. Results: Analysis revealed methodological diversity, with predominance of qualitative approaches (33.3%) and concentration in higher education (35.2%) and teacher training (24.1%). Main applications included problem‐solving support (46.3%), instructional material design (22.2%) and metacognitive scaffolding (14.8%). Benefits encompassed conceptual understanding enhancement (22.2%), metacognitive development (20.4%) and teacher professional growth (20.4%). Critical concerns centred on accuracy and reliability (46.3%), necessitating explicit verification protocols, AI literacy development (31.5%) and strong teacher mediation (27.8%). Conclusion: ChatGPT's educational value is contingent upon instructional design quality rather than tool capabilities. Effective integration demands AI literacy, structured tasks promoting mathematical validation, and pedagogical mediation that preserves disciplinary rigour and student autonomy. Practitioner Notes: What is already known about this topic ○ChatGPT and other large language models have rapidly expanded in educational settings since late 2022, generating both opportunities and concerns about their pedagogical value in mathematics teaching.○Generative AI tools can provide step‐by‐step explanations, generate mathematical examples, and sustain tutorial‐like interactions in natural language.○Unmediated or weakly guided use of ChatGPT can produce plausible but incorrect responses, hallucinations, and may harm learning outcomes, particularly in cognitively demanding mathematical tasks.○Teacher mediation, verification protocols, and AI literacy are considered essential for effective integration of generative AI in mathematics education.○Educational benefits depend more on instructional design quality than on the technological capabilities of the tool itself.What this paper adds ○A systematic synthesis of 54 peer‐reviewed studies (2023‐2025) reveals seven predominant implementation patterns: Direct learning support (24.1%), problem posing/analysis (20.4%), instructional design (16.7%), metacognitive scaffolding (13.0%), secondary studies (11.1%), pedagogical simulation (9.3%), and evaluative approaches (5.6%).○Evidence concentrates in higher education (35.2%) and teacher education (24.1%), with students primarily acting as direct users (37.0%) or critical evaluators (14.8%), while teachers function as designers and mediators (16.7%).○Reported benefits span multiple dimensions: Conceptual understanding (22.2%), metacognitive development (20.4%), teacher professional growth (20.4%), personalized learning (18.5%), operational efficiency (18.5%), and affective improvements (16.7%).○Critical concerns center on accuracy and reliability (46.3% of studies), necessitating explicit verification protocols, alongside risks of overreliance (24.1%) and ethical‐institutional considerations (24.1%).○Effective integration requires operationalized frameworks with explicit components (only 27.3% of studies with frameworks achieved this), revealing a gap between theoretical citation and practical implementation guidance.Implications for practice and/or policy ○Design structurally, not instrumentally: ChatGPT's educational value emerges from deliberate instructional design—defining explicit purposes, timing, task demands (justify, verify, contrast), and transparent validity criteria—rather than from generic adoption or convenience uses.○Prioritize verification as mathematical practice: Given the 46.3% prevalence of accuracy concerns, verification should be positioned as a core component of mathematical activity, not an optional safeguard. This requires explicit rubrics, validation protocols, and tasks that demand justification and correction of AI outputs.○Strengthen AI literacy through disciplinary lenses: Effective AI literacy transcends prompt engineering; it requires developing competencies to analyze, question, and verify outputs against mathematical standards. This includes detecting conceptual errors, biases, and limitations specific to mathematical reasoning.○Implement duration‐sensitive designs: With 44.4% of studies omitting duration details, practitioners must recognize that single‐session experiences capture novelty effects rather than sustained learning. Extended implementations (≥ 6 weeks) with fidelity checks and post‐intervention follow‐ups are necessary to distinguish enthusiasm from pedagogical value.○Develop teacher mediation protocols: Teacher roles must shift from content delivery to validation stewardship—maintaining disciplinary rigor, intervening when errors occur, and ensuring that AI use preserves rather than undermines student autonomy and mathematical thinking.○Establish institutional safeguards: Beyond individual teaching strategies, institutions should develop ethical frameworks addressing academic integrity, privacy, equitable access, and transparent usage policies that protect both students and instructors.○Report implementation components systematically: Future interventions should explicitly document: (a) theoretical frameworks with operational indicators, (b) intervention duration and fidelity measures, (c) psychometric properties of instruments, and (d) task structures that activate mathematical validation—enabling cumulative evidence building and cross‐study comparison. pubtype: Academic Journal doctype: research systematic review tables/charts Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|