AI evidence extraction

Ai literacy and academic integrity in a globalised educational system: A systematic review.

Albert Byiringiro · SJ Education Research Africa · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.51168/x2x7q303

Methodology & findings

Study design

Systematic review following PRISMA 2020 standards.

Sample

N = 29, 3 groups

Primary method

Qualitative thematic synthesis employed for data analysis. Quality appraisal conducted to assess methodologies: conceptual and review articles reviewed for strength of theoretical arguments and relevance to research questions; empirical studies reviewed for methodologies and validity of findings. Two researchers used for screening process to ensure consistency.

Main result

The systematic review found that "AI literacy is an evolving concept that varies widely in definition" and that "issues related to academic integrity, including plagiarism, contract cheating, and authorship, have been exacerbated by generative AI technologies like ChatGPT." Additionally, "AI literacy has been conceptualized as a multidimensional construct that incorporates technical, ethical, and critical dimensions" and the review identified that "the importance of explicit instruction in the ethical use of AI and the critical evaluation of AI outputs is emphasized as a key factor in the mitigation of the misuse of AI in the literature."

Reports effect sizes.

Research paradigm

Interpretive/Critical; mixed-methods synthesis approach

Author conclusions

The authors conclude: "The results show that there is an inconsistent understanding of AI literacy, which involves technical, ethical, and critical aspects." They further state: "Theoretically, this systematic review makes a contribution to the field by conceptualizing AI literacy as a mediator between AI use and academic integrity. Methodologically, it provides a systematic review of this disjointed research area, which is necessary to address calls for greater coherence and rigor in AI in education research." Recommendations include: "First, the integration of AI literacy in higher education curricula, which combines technical skills with ethical considerations and critical evaluation, is recommended. Second, the design of new assessments that focus on reflective, process-based, and authentic learning activities is recommended, which minimizes the scope for AI-assisted cheating. Third, the promotion of inter-institutional and intergovernment collaboration to create academic integrity guidelines that explicitly address AI, considering the diversity of global education systems, is recommended."

Risk of bias

Publication bias toward positive/significant results; Language bias (English-only literature included); Database selection bias (limited to four major databases); Geographic bias (underrepresentation of Global South); Methodological heterogeneity (mix of conceptual and empirical studies); Risk of selective reporting in included studies; Language bias: only English-language publications included; Database bias: limited to Scopus, Web of Science, ERIC, and Google Scholar; Publication bias: positive/significant results more likely published than null/inconclusive results; Study type heterogeneity: inclusion of both conceptual/review literature and empirical studies with varying rigor levels; Geographic bias: dominated by literature from Europe, North America, and Australia with limited Global South representation; Publication bias: literature with significant/positive results more likely to be published; Database coverage bias: limited to specific databases (Scopus, Web of Science, ERIC, Google Scholar); Regional repository bias: may have excluded literature from regional databases; Methodological variability: inclusion of both conceptual and empirical studies with varying rigor levels; Geographic bias: dominance of Global North perspectives (Europe, North America, Australia)

Limitations

The authors state: "the current review has a few limitations, including the fact that it was based on literature reviewed from specific databases, such as Scopus, Web of Science, ERIC, and Google Scholar, which might have left out literature published in other databases or regional repositories
It was also based on literature published in English, which might have left out literature published in other languages, especially where the study was conducted outside English-speaking countries
There was a publication bias, where literature showing significant or positive results was more likely to have been published than literature showing no results or inconclusive results." Additionally, "the inclusion of conceptual and review literature, which might have varied in terms of the level of rigor used in the study."

Open questions raised

The authors identify several gaps: "Cross-national comparative studies are recommended to take into account the diversity of regional policies, technological, and integrity standards, particularly in the underrepresented Global South countries." Additionally, "More attention should be paid to the voices of students and educators, whose experiences with AI tools and academic integrity policies are less explored and less understood." They also note that "longitudinal studies on how AI literacy and integrity standards change over time and their impact on pedagogy are limited." Future research should focus on "designing and conducting empirical and experimental studies to explore the impact of AI literacy interventions on student learning and integrity behaviors over time."
Cross-national comparative studies recommended to account for diversity of regional policies, technological and integrity standards, particularly in underrepresented Global South countries; more attention needed to voices of students and educators whose experiences with AI tools and academic integrity policies are less explored; longitudinal studies on how AI literacy and integrity standards change over time are limited; need for more empirical and experimental studies exploring impact of AI literacy interventions on student learning and integrity behaviors over time; dearth of studies conducted in Global South making it difficult to generalise current models and frameworks
Cross-national comparative studies needed, particularly in underrepresented Global South countries
More attention to voices of students and educators whose experiences with AI tools and academic integrity policies are less explored
Lack of longitudinal studies on how AI literacy and integrity standards change over time
Limited empirical studies on student learning and behavior related to AI

Extracted from: pdfAgreement 66%

Explore related topics

AI Literacy Training Research Integrity