Analyzing ChatGPT and Generative AI in Education

Table of Links

Abstract and Introduction

History of Using AI in Education

Research Methodology

Literature Review

Summary

Conclusion and References

3. Research Methodology

This section discusses the research methodology we followed to develop search, filter, and select the related work on using generative AI models in general and ChatGPT in particular in different educational settings.

3.1. Search approach

In this review, the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) (Moher et al., 2009; Knobloch et al., 2011) statement was followed to select relevant articles. The search for articles started on June 25, 2023, until July 10. The version of ChatGPT discussed in the articles was the original release on November 30, 2022 (i.e. ChatGPT 3.5). The search keywords (”ChatGPT”, ”Education”, ”Generative AI”, ”Chatpot”) were used to search for related publications in different databases including ACM, IEEE Xplore, Scopus, and Web of Science. Articles that mentioned these keywords in their title, abstract, or keywords were selected. The publication period was filtered to be after 2022. Despite searching multiple databases, the number of relevant articles found was limited. To ensure a comprehensive search, Google Scholar was used to conduct a title search for the term ”ChatGPT” within the same search period. This additional search method allowed for the identification of relevant articles that may not have been captured in the initial database search.

3.2. Paper Selection Criteria

In this review, academic articles published between November, 2022, and July, 2023, were included. The review included only advanced online publications published in journals indexed by Scopus as Q1 or Q2 journals. The focus of the review was on articles that discussed the application of generative AI models (mainly ChatGPT) in education, without any specific constraints on educational contexts. Identified literature reviews were used as background references and were excluded from the synthesis to avoid redundancy in references. Moreover, Only articles written in English were included in this rapid review, ensuring a consistent language focus for the analysis.