Where is AIED Headed? Key Topics and Emerging Frontiers (2020-2024)
Shihui Feng, Huilin Zhang, Dragan Gašević · 2025
Read Where is AIED Headed? Key Topics and Emerging Frontiers (2020-2024) by Shihui Feng, Huilin Zhang, Dragan Gašević (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-20971.txt.
Introduction
Artificial intelligence in education has been remade by generative AI in just a few years. Feng, Zhang, and Gašević map the field's knowledge structure by analyzing 2,398 articles published from 2020 to 2024 in eight core venues.
Using network analysis of how keywords appear together, they trace the field's stable clusters and its rising edges. The durable core is technical: intelligent tutoring systems, learning analytics, and natural language processing. Four frontiers are emerging on top of it: large language models, generative AI itself, multimodal learning analytics, and human-AI collaboration. Work on generative tools clusters around personalization, self-regulated learning, feedback, assessment, and ethics. The authors read the trend as a turn toward systems that adapt with the learner rather than merely at the learner.
Catalog record
| CREATOR | Shihui Feng, Huilin Zhang, Dragan Gašević |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 63 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-20971
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Where is AIED Headed? Key Topics and Emerging Frontiers (2020-2024) Shihui Feng, Huilin Zhang, Dragan Gašević Abstract In this study, we analyze 2,398 research articles published between 2020 and 2024 across eight core venues related to the field of Artificial Intelligence in Education (AIED). Using a three-step knowledge co-occurrence network analysis, we analyze the knowledge structure of the field, the evolving knowledge clusters, and the emerging frontiers. Our findings reveal that AIED research remains strongly technically focused, with sustained themes such as intelligent tutoring systems, learning analytics, and natural language processing, alongside rising interest in large language models (LLMs) and generative artificial intelligence (GenAI). By tracking the bridging keywords over the past five years, we identify four emerging frontiers in AIED—LLMs, GenAI, multimodal learning analytics, and human-AI collaboration . The current research interests in GenAI are centered around GAI-driven personalization, self-regulated learning, feedback, assessment, motivation, and ethics.The key research interests and emerging frontiers in AIED reflect a growing emphasis on co-adaptive, human-centered AI for education . This study provides the first large-scale field-level mapping of AIED’s transformation in the GenAI era and sheds light on the future research development and educational practices. Keywords: AIED, Generative AI, Human-AI collaboration, large language models, multimodal learning analytics, keyword co-occurrence networks I Introduction The field of Artificial Intelligence in Education (AIED) has undergone rapid development in recent years, driven by advances in machine learning, natural language processing, and the increasing integration of AI into educational platforms. AIED as an interdisciplinary research field transcends disciplinary boundaries and integrates methods and tools from computer science and information science to solve educational problems (Pinkwart, 2016; Feng & Law, 2021; Hwang et al., 2020). Its applications span diverse educational scenarios—from formal classroom instruction to informal self-directed learning—enabling dynamic, personalized knowledge dissemination across contexts while supporting process-oriented assessment and real-time feedback (Guan et al., 2024; Ayemowa et al., 2024; Dai et al., 2024; Hwang et al., 2020). The release of GPT-3 by OpenAI in 2020 marked a significant progress in generative artificial intelligence (GenAI), showcasing unparalleled capabilities in language understanding and generation (Aryadoust et al., 2024). This advancement made the generation of human-like content feasible, leading to widespread applications across various fields, including education (Watson & Shi, 2024). The advent of large language models (LLMs) has expanded AI applications in educational contexts beyond traditional tasks such as automated grading and intelligent question answering to advanced functions like generating personalized lesson plans, delivering real-time learning recommendations, and creating immersive experiences for language acquisition (Alaqlobi et al., 2024; Salih et al., 2024; Wang et al., 2025). These advancements are redefining the research paradigms and practical applications of AIED, paving the way for new opportunities and possibilities in the future of education. Given the changes happening in the field, tracking shifts in research focus of AIED—through the lens of keyword co-occurrence networks—offers valuable insights into emerging priorities, educational applications, and future research directions. Feng & Law (2021) analyzed the keyword co-occurrence network of 1,830 articles in the AIED field from 2010 to 2019. They identified that the sustained themes in AIED were intelligent tutoring systems and massive open online courses, and highlighted the emerging topics including neural networks, deep learning, eye tracking, and personalized learning during that period. In retrospect, the development of neural networks and deep learning has enabled the success of GenAI, which provides foundation models trained on large-scale dataset and propels AI applications in education to grow rapidly in recent years. Built on this previous work, in the current study, we analyzed the keyword co-occurrence network of 2,398 articles from eight key venues related to AIED for the period 2020-2024, to identify the recent key research focus and uncover the intrinsic connections among diverse interests within the field. Specifically, this study was guided by the following three research questions: RQ1: What is the knowledge structure of AIED between 2020 and 2024 (five years)? RQ2: What are the key knowledge clusters of AIED between 2020 and 2024? RQ3: What are the emerging frontiers in AIED? II Related Work With the ongoing expansion of AIED research, comprehensive reviews are needed for elucidating the key topics in the field and offering directions for future studies. Several review articles have summarized major research topics in the field of AIED. For instance, Zhai et al. (2021) conducted a content analysis of 100 studies published between 2010 and 2020, identifying major research dimensions such as classification, recommendation, deep learning, real-time feedback, and adaptive learning. They also outlined four critical future research directions: the Internet of Things, collective intelligence, deep learning, and neuroscience. Mustafa et al. (2024) reviewed 143 reviews in the AIED field, using the technology-enhanced learning model to code the literature. Their findings indicate that AI applications in education primarily target higher education, teacher support, and student learning, while receiving less attention for administrators, school leaders, and special education. In addition, there are reviews focusing on specific tools or application scenarios within AIED, including evaluating the features and applications of AI tools such as intelligent tutoring systems (ITS), automated scoring systems, recommendation systems, and learning analytics. Mousavinasab et al. (2018) examined 53 papers published between 2007 and 2017, systematically summarizing variations and technological foundations of ITS developed for diverse educational contexts. Huawei and Aryadoust (2023) analyzed 105 articles published before May 2020, assessing the application contexts, effectiveness, and limitations of automated scoring systems. Additionally, Additionally, some studies employed meta-analyses to evaluate the impact of AI tools, revealing significant effects on learning outcomes, medium improvements in higher-order thinking (HOT) skills, and smaller but notable influences on learning perceptions (Alemdag, 2023; Dibek et al., 2024; Zheng et al., 2021). The application scenarios of AIED are also broad, encompassing medical education, STEM education, language teaching, K-12 education, mathematics education, and higher education (Hwang & Tu, 2021; Xu & Ouyang, 2022; Zafari et al., 2022). For example, some researchers have used narrative or scoping reviews to explore the forms of AI applications, student experiences, and limitations in medical education (Preiksaitis & Rose, 2023). Xu and Ouyang (2022) conducted a systematic review of 63 empirical AI-STEM studies from 2011 to 2021, summarizing six categories of AI applications and illustrating the relationships between AI application types and other elements (e.g., information, subjects, media, and environments) within AI-STEM systems. Similarly, Zafari et al. (2022) performed a systematic review of 210 articles and conference papers indexed in the Web of Science and Scopus databases from 2011 to 2021. They categorized AI applications in K-12 education into four dimensions: student performance, teaching, selection and behavior tasks, and others. A few studies employed a bibliometric approach to analyze AIED literature, aiming to have a comprehensive description of the research field. For instance, Delen et al. (2024) analyzed 4,673 publications from 1975 to 2023 involving 29 AIED-related keywords, providing a multidimensional visualization of annual publication patterns, disciplinary distribution, and leading authorship within the field. Responding to the recent rise of GenAI, Polat et al. (2024) conducted a bibliometric study of 212 Scopus-indexed publications up to July 2023, identified via the keywords “ChatGPT” and “education.” The study revealed ChatGPT’s exponential growth in education with focuses on computational intelligence, human learning, language processing, and conversational agents. Bozkurt et al. (2021) applied social network analysis and text-mining techniques