The Global Landscape of Generative AI in Higher Education
A new article by an international team of researchers, including Evgeniy Terentev, Director of the HSE Institute of Education, provides a comprehensive analysis of over 4,000 academic publications on Generative AI in higher education from 2022 to 2024.
Titled 'Mapping the Generative AI Research in Higher Education: 2022–2024 Insights', the article was published in the prestigious journal Higher Education Quarterly in November. The research was conducted by a collaborative team including Isak Frumin and Daria Platonova (Constructor University, Germany), Anton Vorochkov (Autonomous University of Madrid, Spain), Margarita Kiryushina (Technion Israel Institute of Technology, Israel) and Evgeniy Terentev (HSE University, Russia).
With the release of the first version of ChatGPT on November 30, 2022 the general interest in AI was transformed into the explosive growth of studies of generative AI and large language models. Our estimate shows a threefold increase in academic publications between 2023 and 2024.
As artificial intelligence is constantly evolving, and the impact of this evolution on higher education is becoming increasingly relevant, scholars suggest adopting a two-perspective approach to this issue.
Firstly, they analyse the current landscape of academic publications in this field to provide an overview of this rapidly developing area. The main goal of this article is to contribute to the growing body of research on the role of GenAI in higher education.
Secondly, they study the academic community's reaction to the emergence of GenAI. The focus is on higher education, with the aim of identifying insights for further discussion of its development.
As a result, the international team was able to analyse 4,145 research publications collected from the Scopus database between 2022 and 2024. In order to capture the dynamic and rapidly evolving discourse surrounding GenAI in higher education, the scholars adopted an inclusive approach to publication formats and languages. This included traditional peer-reviewed articles, conference proceedings and preprints, as well as publications in various languages.
A comparison with the broader field of higher education research, also based on Scopus data, highlights the distinct communication dynamics of AI-focused studies.
In addition to identifying publication dynamics in recent years (increasing from 215 in December 2022 to 2,679 in November 2024), researchers revealed trends in publication formats. Although journal articles remain the most common format, accounting for 54% of publications, conference papers have also shown moderate growth. The researchers noted that the emphasis on conference-based communication highlights the rapid pace of AI advances. AI scholars tend to choose venues that facilitate the rapid exchange of results, reflecting the dynamic nature of the field.
This distribution indicates a clear shift from studying the technical capabilities of GenAI to investigating their practical implementation in education, ethical implications, and integration into existing educational frameworks.
Regarding current research topics related to GenAI, it was found that the focus is primarily on teaching and learning methods (36%) and ethical considerations (27%), particularly in relation to pedagogical implementations and issues of academic integrity. This focus on ethics reflects the growing concern about the responsible use of GenAI in higher education.
Looking ahead, we anticipate the emergence of more sophisticated research methodologies that can better capture the complexities of AI implementation in educational settings.
Although the bibliometric study provided a comprehensive overview of the research field, the analysis highlighted the need for a more detailed, qualitative examination of individual cases and contexts. Among other things, scholars suggest that future research should focus on national and institutional studies to understand how AI educational solutions can be adapted to different systems. Longitudinal and randomised studies would also be useful for tracking effects over time and establishing causal links.