Reviewing Researches on Artificial Intelligence—Driven Education and Teaching from 2014 to 2023: A CiteSpace-based Visual Analysis

Xin HuangFrontiers in Educational Research 2024 · DOI

TL;DR

This paper provides a comprehensive bibliometric "map" of AI-driven education research from 2014 to 2023, identifying three distinct developmental stages: germination, boom, and stabilization. By analyzing over 1,300 high-quality papers, the study reveals a significant shift from basic theoretical exploration toward data-driven applications like learning analytics and the recent emergence of large language models.

What Problem Does This Paper Address?

As Artificial Intelligence (AI) moves from science fiction to the classroom, the sheer volume of academic literature has become overwhelming. Researchers and policymakers often struggle to identify which themes are truly "hot," who the leading experts are, and how the conversation has evolved over time.

This paper addresses this "information overload" by providing a structured, visual analysis of the research landscape in China. It seeks to answer: Which institutions are leading the charge? How have research priorities shifted from 2014 to today? And what are the "frontier" topics that will define the next decade of AI in education?

Methods

The author employed bibliometrics and knowledge mapping using CiteSpace, a specialized visual analytics tool. Key techniques included:

  • Cluster Analysis: Grouping related research keywords to identify dominant themes.
  • Centrality Analysis: A mathematical measure used to identify "bridge" institutions or keywords that connect different sub-fields, indicating high influence.
  • Keyword Burst Analysis: Detecting terms that experienced a sudden surge in frequency, which signals the "frontier" of the field.
  • Price’s Law: The author used the formula $M = 0.749\sqrt{N_{max}}$ to quantitatively identify "core authors" who contribute the most to the discipline.

Datasets

The study utilized a dataset of 1,364 papers retrieved from the China National Knowledge Infrastructure (CNKI) database.

  • Selection Criteria: The search was limited to CSSCI (Chinese Social Sciences Citation Index) journals to ensure high data quality and academic rigor.
  • Timeframe: Literature published between 2014 and November 25, 2023.
  • Keywords: The search combined "Artificial Intelligence" with "Education," "Teaching," and "Talent Cultivation."

Key Insights and Ideas

  • The Three-Act Evolution: The field didn't grow linearly. It had a Germination Period (2014–2016) with low volume, a Boom Period (2017–2019) characterized by a massive spike in interest, and a current Stabilization Period (2020–2023) where research has become more mature and focused on sustainable implementation.
  • Institutional Powerhouses: Research is heavily concentrated in "Normal" (teacher-training) universities. East China Normal University and Beijing Normal University emerged as the two most influential hubs, acting as the "central nervous system" for collaboration networks.
  • From Theory to Data: Early research was preoccupied with defining what AI could do. The focus has since shifted toward Learning Analytics and Big Data, indicating a move toward evidence-based interventions where student data is used to personalize learning paths in real-time.
  • The "Human" Rebound: Interestingly, as AI technology matures, research into the "Teacher Role" and "Educational Ethics" has seen a "burst" in popularity. This suggests that the more capable AI becomes, the more we worry about the philosophical and social implications of human-machine collaboration.

Key Results

  • Core Authors: Using Price's Law, the study identified 37 "core authors," with Xiaoqing Gu leading the field with 31 publications.
  • High-Impact Keywords: "Artificial Intelligence" (825 occurrences) and "Talent Cultivation" (66 occurrences) are the most frequent, but "Deep Learning" showed the highest burst strength (5.03) between 2017 and 2018, marking it as a pivotal technological turning point.
  • Geographic Concentration: Collaboration clusters are tightly tied to geography, with major research groups centered in technologically advanced regions like Beijing, Shanghai, and Jiangsu.

Limitations and Future Work

  • Geographic Bias: The study focuses exclusively on the CNKI database. While this offers a deep dive into Chinese academia, it does not incorporate international trends from databases like Web of Science or Scopus.
  • The Generative Gap: While the paper identifies "ChatGPT" and "Large Models" as emerging frontiers, the data only captures the very beginning of this wave. The author suggests that future research must specifically address how generative AI disrupts traditional assessment and teaching methodologies.
  • Siloed Research: Despite the presence of core groups, a large number of researchers still work independently. The paper calls for more cross-disciplinary collaboration to bridge the gap between computer science and educational theory.

Summary updated

Abstract

In the current digital era, the field of education is undergoing unprecedented changes. Artificial Intelligence (AI), as a revolutionary technology, has demonstrated its unique advantages in various fields such as healthcare, finance, and manufacturing. To explore the application of AI in education and teaching, as well as its evolving trends, we analyzed 1364 CSSCI papers from the China National Knowledge Infrastructure (CNKI) database from 2014 to 2023. Employing bibliometrics and knowledge mapping methods, we established a detailed view of the development trends and research hotspots in AI-driven education and teaching research. The Citespace software was employed for visual analysis of critical data such as authors, institutions, and keywords, including the collaboration networks of authors and institutions, keyword frequency and centrality, and the evolution of research topics. Our analysis revealed that AI-driven education and teaching research from 2014 to 2023 can be divided into three developmental stages: nascent (2014–2016), rapid development (2017–2019), and stable development (2020–2023). What's more, we identified close cooperation among core institutions and authors, determined multiple hotspot topics in AI within education and teaching, and revealed the developmental trajectory of these topics.
BibTeX
@inproceedings{huang2024reviewing,
  author = {Xin Huang},
  title = {{Reviewing Researches on Artificial Intelligence—Driven Education and Teaching from 2014 to 2023: A CiteSpace-based Visual Analysis}},
  booktitle = {Frontiers in Educational Research},
  year = {2024},
  doi = {10.25236/fer.2024.070102},
}
Figure