Artificial intelligence in mathematics education: A systematic literature review

Artificial intelligence in mathematics education: A systematic literature review Mohamed Hidayat Suhaizi Sabri Mahmud Baharuddin

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

Artificial intelligence in mathematics education: A systematic literature review 2022 International Electronic Journal of Mathematics Education 10.29333/iejme/12132 mohamed2022-artificial-intelligence-in-mathema mohamed2022-artificial-intelligence-in-mathema.txt, the project's staged copy 136eebcfdec857574626085d24034faca608d5b6 open-access education

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them existence Quotation withheld because this record is not cleared for redistribution. artificial intelligence in mathematics education about this quote: This quote ends with the stray word 'A', the first word of the next sentence. It is not trimmed, because the quote without it occurs more than once in the paper's text and could not be tied to one place. The claim is carried by the sentence before 'A'. Quotation withheld because this record is not cleared for redistribution. systematic-review systematic-review Quotation withheld because this record is not cleared for redistribution. mechanisms the high maintenance cost of AI equipment restricts who can run such studies, which is why the reviewed research clusters in developed countries and is absent from Southeast Asia Quotation withheld because this record is not cleared for redistribution. Only a small number of eligible articles exist, so broad conclusions cannot be drawn Quotation withheld because this record is not cleared for redistribution. AI is rarely investigated in mathematics education, limiting the results obtained Quotation withheld because this record is not cleared for redistribution. De-duplication across search engines reduced the number of relevant studies acquired Quotation withheld because this record is not cleared for redistribution. Only journal articles were included; proceedings and books were excluded Quotation withheld because this record is not cleared for redistribution. The review's central term is stated to have no agreed definition, and the review adopts none. what this limits: any count reported in this review: the authors state there are diverse concepts of AI in the literature and describe defining AI as a problem, without settling on one definition for the review Quotation withheld because this record is not cleared for redistribution. The population is defined by a list of search synonyms rather than by a concept. what this limits: the scope of what this review counts as AI: the search treated artificial intelligence, AI, robotics, machine learning, neural networks, fuzzy logic, natural language processing and expert systems as synonyms and alternative terms for one main search term Quotation withheld because this record is not cleared for redistribution. The three article counts printed for the search stage are inconsistent with each other. what this limits: the PRISMA flow reported here: 864 results from the databases plus 20 from other sources are stated to give a total of 885, and the printed numbers do not sum Quotation withheld because this record is not cleared for redistribution. The corpus is restricted by language and by a five-year window. what this limits: any claim about how much AI research exists in mathematics education: only English-language journal articles published between 2017 and 2021 were considered Quotation withheld because this record is not cleared for redistribution. A PROCEDURAL SAFEGUARD ON A SYSTEMATIC REVIEW: every one of the 845 articles read through title, abstract, methodology, results and discussion against the stated selection criteria, rather than screened on title and abstract alone eliminates: articles that match the search terms but not the research question - a review screened on titles inherits whatever the titles happen to say, and the synthesis then describes the search rather than the literature · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. artificial intelligence artificial intelligence Quotation withheld because this record is not cleared for redistribution. machine intelligence machine intelligence Quotation withheld because this record is not cleared for redistribution. expert systems expert systems Quotation withheld because this record is not cleared for redistribution. intelligent tutoring system intelligent tutoring system Quotation withheld because this record is not cleared for redistribution. computer program computer program Quotation withheld because this record is not cleared for redistribution. AI - THREE CHARACTERISATIONS AND NO RULING: the review reports that there are diverse concepts of AI in the literature and that defining it is a problem, then quotes a confluence of technologies, computers that do cognitive functions, and a process that produces human intelligence through machines - three sources, one review, and no statement of which the review is using Quotation withheld because this record is not cleared for redistribution. AI, OPERATIONALLY - A LIST OF EIGHT SEARCH STRINGS: what the review actually counts is papers matching artificial intelligence, AI, robotics, machine learning, neural networks, fuzzy logic, natural language processing or expert systems, so a table of search synonyms is doing the work that none of the three definitions does Quotation withheld because this record is not cleared for redistribution. machine intelligence - THE ALTERNATIVE NAME, GIVEN IN PASSING: the review introduces the problem of defining AI as the problem of defining artificiality's parameters, and names the field 'machine intelligence' in the same clause, attributing it to Poole et al. 1998 Quotation withheld because this record is not cleared for redistribution. builds-on → Ouyang and Jiao (2021) conceptual framework of AI-directed / AI-supported / AI-empowered learning Quotation withheld because this record is not cleared for redistribution. builds-on → Moher et al. (2009) PRISMA flow diagram Quotation withheld because this record is not cleared for redistribution. replicates → Zhong and Xia (2020) systematic review of robotics in mathematics learning Quotation withheld because this record is not cleared for redistribution. intelligent tutoring system —is-a→ computer program intelligent tutoring system computer program strength of the relation: 1 Quotation withheld because this record is not cleared for redistribution. Robotics was the most common AI approach in the 20 reviewed mathematics-education studies, accounting for 45% (n=9) of them. Quotation withheld because this record is not cleared for redistribution. The USA contributed the most reviewed studies (n=3). Quotation withheld because this record is not cleared for redistribution. Quantitative designs were the most used methodology among the reviewed studies, at 40% (n=8). Quotation withheld because this record is not cleared for redistribution. 2021 was the peak publication year for AI-in-mathematics-education articles in the review window, with 30% of the corpus. Quotation withheld because this record is not cleared for redistribution. Effectiveness of AI was the most studied theme, in 60% (n=12) of the reviewed papers. Quotation withheld because this record is not cleared for redistribution. 834 articles were rejected for not fully explaining AI in mathematics education or not clearly reporting findings. Quotation withheld because this record is not cleared for redistribution. artificial intelligence (AI) artificial intelligence (AI) Quotation withheld because this record is not cleared for redistribution. systematic literature review (SLR) systematic literature review (SLR) Quotation withheld because this record is not cleared for redistribution. PRISMA PRISMA Quotation withheld because this record is not cleared for redistribution. intelligent tutoring system intelligent tutoring system Quotation withheld because this record is not cleared for redistribution. research methodology research methodology Quotation withheld because this record is not cleared for redistribution. 45%, n=9 — share of reviewed studies using a robotics approach to AI Quotation withheld because this record is not cleared for redistribution. 40%, n=8 — share of reviewed studies using quantitative methods Quotation withheld because this record is not cleared for redistribution. 60%, n=12 — share of reviewed studies whose theme was effectiveness of AI Quotation withheld because this record is not cleared for redistribution. 30% — share of the reviewed corpus published in 2021, the peak year Quotation withheld because this record is not cleared for redistribution. 864 — records retrieved by the database searches, plus 20 from other sources Quotation withheld because this record is not cleared for redistribution. 834 — articles rejected at the eligibility stage Quotation withheld because this record is not cleared for redistribution. 20 — articles retained for the final synthesis Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.29333/iejme/12132.