Abstract:

Artificial Intelligence (AI) is becoming one of the most influential emerging technologies of the moment. In the context of education, the proliferation of chatbots based on generative artificial intelligence, such as ChatGPT, has revolutionised the academic system. This technology generates text and can be used by students to complete academic tasks, by professors to create content and by institutional managers to develop administrative tasks. However, the use of AI tools also implies negative issues related to plagiarism, transparency, privacy, and false information that these technologies may provide. These issues classify their impact in terms of ethics, critical thinking, and creativity. Thus, their use in education is attracting the attention of governments, academics, and institutions. It is since the Covid-19 pandemic that these technologies have been promoted, considering the situation of isolation in this period and the promotion of digitalisation required by institutions. The aim of this study is to shed light on the state of the art in the use of AI in higher education, to identify research trends to understand the conceptual structure and to drive future research in this field. This paper follows a bibliometric approach. with a total sample of 340 papers retrieved from the Web of Science (WoS) database. The results are mapped using VOSviewer software A co-occurrence analysis is performed to identify the research hotspots. Four main clusters are obtained, focusing on academic performance and technology acceptance, underscored keywords such as 'machine learning' and 'robotics’.

Other techniques have been used, including citation analysis, considering the most cited papers to determine the most influential lines of research. The main themes are related to virtual laboratories, student engagement and prediction of performance and satisfaction, the use of data mining and robotics. Most of these papers -almost sixty percent- are published in the journal Computers & Education, oriented to computer science. The results indicate that this field has been mainly retained attraction on computing and engineering degrees. Thus, there is a need for more emphasis on business and management, psychology, or marketing related subjects, among others, and the use of AI in teaching in these domains. Most of the articles are empirical and some of them are case studies. Also, the most prolific journals and countries -to identify where this domain is receiving more attention- have been considered. The journals Education and Information Technologies, International Journal of Engineering Education and IEEE Transactions on Education stand out as the most productive in terms of the total number of articles published. According to the authors, those from the United States, China and Spain form the top core. In addition, future research directions were suggested in relation to the use of ChatGPT in higher education, new strategies to improve student engagement, as well as comparisons between different regions and higher education institutions, and how to measure performance in relation to the application of AI.

Keywords:

Artificial intelligence, higher education, ChatGPT, bibliometric analysis, state-of-the-art.