Source-linked AI summary
A Systematic Review of Generative AI for Teaching and Learning Practice
Bayode Ogunleye, Kudirat Ibilola Zakariyyah, Oluwaseun Ajao, Olakunle Olayinka, Hemlata Sharma
TL;DR
Unclear guidance and limited systematic evidence make GenAI’s effective use in higher-education teaching and learning uncertain. The paper reviews Scopus-indexed research using PRISMA, bibliometric indicators, and topic modelling, identifying publication patterns and major themes while highlighting curriculum, assessment, ethical, and interdisciplinary research needs.
Problem
There are no agreed higher-education guidelines for GenAI use, and limited systematic reviews leave its effective teaching and learning applications unclear.
Method
The study systematically reviewed Scopus-indexed higher-education GenAI literature using PRISMA, bibliometric indicators, and latent Dirichlet allocation topic modelling.
Results
The review found that journal articles comprised 72% of documents and identified trends spanning academic integrity, assessment, chatbots, ethics, intelligent tutoring, and GenAI applications.
Takeaways & Limitations
The findings support further curriculum and assessment research, interdisciplinary collaboration, and stakeholder understanding to inform GenAI guidelines, frameworks, and policies.
Takeaways & Limitations
Future evidence should expand representation through non-English publications and use longitudinal studies to monitor changing GenAI research trends.
Abstract
from arXiv · showhide
The use of generative artificial intelligence (GenAI) in academia is a subjective and hotly debated topic. Currently, there are no agreed guidelines towards the usage of GenAI systems in higher education (HE) and, thus, it is still unclear how to make effective use of the technology for teaching and learning practice. This paper provides an overview of the current state of research on GenAI for teaching and learning in HE. To this end, this study conducted a systematic review of relevant studies indexed by Scopus, using the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. The search criteria revealed a total of 625 research papers, of which 355 met the final inclusion criteria. The findings from the review showed the current state and the future trends in documents, citations, document sources/authors, keywords, and co-authorship. The research gaps identified suggest that while some authors have looked at understanding the detection of AI-generated text, it may be beneficial to understand how GenAI can be incorporated into supporting the educational curriculum for assessments, teaching, and learning delivery. Furthermore, there is a need for additional interdisciplinary, multidimensional studies in HE through collaboration. This will strengthen the awareness and understanding of students, tutors, and other stakeholders, which will be instrumental in formulating guidelines, frameworks, and policies for GenAI usage.
1. Introduction
GenAI’s rapid uptake in higher education brings both educational promise and concerns, while limited systematic evidence leaves its effective use for teaching and learning unclear. This review therefore maps the field’s research state, themes, and future directions.
- GenAI tools have been used to find literature, answer student questions, support code implementation, and generate exercises in higher-education teaching and learning.
- The literature raises concerns about hallucinations, bias, ethical and privacy issues, accidental plagiarism, and academic integrity.
- The review aims to overview GenAI research in higher education and synthesize its potential uses, impacts, and ethical issues.
- The study examines field productivity and influence, including journals, citations, authorship, and geographical distribution.
- It also investigates emerging trends and core themes in the existing literature.
- The review contributes an overview of research progression, a synthesis of GenAI’s potential and limitations, and identified gaps for future investigation.
2. Methodology
The study used a PRISMA-guided Scopus review of recent higher-education GenAI literature, combining bibliometric analysis with topic modelling. Screening reliability was assessed through independent coding and Cohen’s kappa.
- The review examined conference proceedings and journal papers from the previous seven years using PRISMA guidelines and Scopus metadata.
- The initial Scopus search produced 625 papers using terms related to generative AI, assessment, higher education, teaching, and learning.
- Two researchers independently screened 20 randomly selected documents, with a third author resolving disagreements during quality assessment.
- The inter-rater reliability assessment produced a Cohen’s kappa value of 0.659.
- Bibliometric indicators covered publications, citations, cited sources and authors, co-authorship, and term co-occurrence.
- Latent Dirichlet allocation topic modelling was evaluated using coherence scores, perplexity, and human interpretation to identify latent themes.
3. Results and Discussion
The reviewed literature expanded sharply in 2023, with journal articles comprising most documents. The publication surge followed ChatGPT’s public release and reflected growing interest in GenAI research.
- The analysis was organized into bibliometric indicators and topic-modelling results.
- 72% of extracted documents were journal articles, compared with 28% conference proceedings.
- Publications increased from 38 to 273 in 2023, a jump associated with ChatGPT’s public launch.
Document Types
The review maps GenAI research in higher education across publication patterns, influential sources and authors, collaboration, keywords, and topic clusters. The literature spans educational applications, support systems, intelligent tutoring, assessment, writing, technical development, and associated ethical and inclusion concerns.
- Citation patterns: The ten most cited sources split evenly between education-focused and technology-focused journals.The Journal of Applied Learning and Teaching led with 301 citations, followed by the International Journal of Information Management with 291.
- Collaboration: The United States was among the most relevant countries in the co-authorship analysis, with 124 papers and 1,598 citations.The country analysis included 35 countries that had at least three papers.
- Keyword trends: Keyword co-occurrence identified artificial intelligence and ChatGPT as the most frequent keywords, with 141 and 126 occurrences, respectively.Keywords associated with 2023 included academic integrity, assessment, ethics, higher education, and prompt engineering.
4. Conclusions
The review maps GenAI research in higher education through publication, citation, authorship, keyword, and thematic patterns, while identifying gaps for curriculum integration, interdisciplinary work, and inclusive evidence.
- Research landscape: 72% of publications were journal articles and 28% were conference papers, with exponential publication growth occurring in 2023.Yogesh K. Dwivedi was the most cited author, and the cited article was the most cited paper.
- Research themes: Keyword and topic analyses identified academic integrity, assessment, ChatGPT, ethics, prompt engineering, and related GenAI applications as central research trends.Core themes also included support systems, bias and inclusion, intelligent tutoring, exam performance, writing, and ethical or regulatory considerations.
- Implications: The review uses bibliometric analysis and topic modelling to provide a holistic view of GenAI’s potential, effectiveness, limitations, and research directions in higher education.The authors present topic modelling as a complement to keyword co-occurrence analysis for distilling latent themes.
- Limitations: The English-language scope may have excluded important studies and non-English topic trends, while medical-education research was disproportionate to research in other higher-education disciplines.The authors recommend broader journal representation, non-English publications, and cross-disciplinary studies.
- Future directions: Future research should examine GenAI across disciplines, update curricula and assessments, evaluate AI-content detectors, and develop guidelines through interdisciplinary collaboration.The review also calls for stakeholder input and research synthesis from students and academic tutors.