Next Article in Journal
Improving Sentiment Analysis Performance on Imbalanced Moroccan Dialect Datasets Using Resample and Feature Extraction Techniques
Next Article in Special Issue
Transitioning from MLOps to LLMOps: Navigating the Unique Challenges of Large Language Models
Previous Article in Journal
Supporting Knowledge Transfer on Functional Significance of Forest Biodiversity
Previous Article in Special Issue
Design and Implementation of an Intelligent Web Service Agent Based on Seq2Seq and Website Crawler
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Information
Volume 16
Issue 1
10.3390/info16010038
information-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Bibliometric Analysis on ChatGPT Research with CiteSpace

by
Dongyan Nan
1,†,
Xiangying Zhao
2,†,
Chaomei Chen
3,
Seungjong Sun
4,5,
Kyeo Re Lee
6 and
Jang Hyun Kim
2,4,5,*
1
School of Business, Macau University of Science and Technology, Macau 999078, China
2
Department of Interaction Science, Sungkyunkwan University, Seoul 03063, Republic of Korea
3
College of Computing and Informatics, Drexel University, Philadelphia, PA 19104, USA
4
Department of Human-Artificial Intelligence Interaction, Sungkyunkwan University, Seoul 03063, Republic of Korea
5
Department of Applied Artificial Intelligence, Sungkyunkwan University, Seoul 03063, Republic of Korea
6
Center for Creative Convergence Education, Hanyang University ERICA, Ansan 15588, Republic of Korea
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Information 2025, 16(1), 38; https://doi.org/10.3390/info16010038
Submission received: 8 October 2024 / Revised: 1 December 2024 / Accepted: 13 December 2024 / Published: 9 January 2025
(This article belongs to the Special Issue Natural Language Processing (NLP) with Applications and Natural Language Understanding (NLU))
Browse Figures
Versions Notes

Abstract

:
ChatGPT is a generative artificial intelligence (AI) based chatbot developed by OpenAI and has attracted great attention since its launch in late 2022. This study aims to provide an overview of ChatGPT research through a CiteSpace-based bibliometric analysis. We collected 2465 published articles related to ChatGPT from the Web of Science. The main forces in ChatGPT research were identified by examining productive researchers, institutions, and countries/regions. Moreover, we performed co-authorship network analysis at the levels of author and country/region. Additionally, we conducted a co-citation analysis to identify impactful researchers, journals/sources, and literature in the ChatGPT field and performed a cluster analysis to identify the primary themes in this field. The key findings of this study are as follows. First, we found that the most productive researcher, institution, and country in ChatGPT research are Ishith Seth/Himel Mondal, Stanford University, and the United States, respectively. Second, highly cited researchers in this field are Tiffany H. Kung, Tom Brown, and Malik Sallam. Third, impactable sources/journals in this area are ARXIV, Nature, and Cureus Journal of Medical Science. Fourth, the most impactful work was published by Kung et al., who demonstrated that ChatGPT can potentially support medical education. Fifth, the overall author-based collaboration network consists of several isolated sub-networks, which indicates that the authors work in small groups and lack communication. Sixth, United Kingdom, India, and Spain had a high degree of betweenness centrality, which means that they play significant roles in the country/region-based collaboration network. Seventh, the major themes in the ChatGPT area were “data processing using ChatGPT”, “exploring user behavioral intention of ChatGPT”, and “applying ChatGPT for differential diagnosis”. Overall, we believe that our findings will help scholars and stakeholders understand the academic development of ChatGPT.

1. Introduction

ChatGPT, a generative artificial intelligence (AI) chatbot developed by OpenAI, has attracted great attention since its launch in late 2022 [1,2]. ChatGPT is a large language model (LLM) equipped with billions of parameters and trained on datasets comprising tens or hundreds of billions of tokens [3]. Its extensive scale allows it to effectively process, generate, and reason with natural human languages [4,5,6]. The most recent base model, GPT-4, has showcased its ability to attain an impressive score, within the top 10%, in bar exams [7]. This exceptional performance has led to its widespread adoption by numerous users [8]. Additionally, the academic community has shown increasing interest in employing ChatGPT as both a research subject and tool [9,10,11].
Despite the explosive development of ChatGPT, there are rare studies aiming to provide an overview of ChatGPT research. To address this research gap, our study attempts to conduct a bibliometric analysis of ChatGPT research. We believe that our research can contribute to accelerating the development of ChatGPT-related academic and industrial fields. Because the outcomes of a bibliometric analysis of a particular field generally help scholars and stakeholders to quickly understand its development status and overall knowledge structure [12]. For example, for policymakers and industry leaders, bibliometric trends also can guide funding allocation and R&D strategies by identifying research hotspots.
Specifically, based on the guidelines for bibliometric analysis [12,13], we aimed to evaluate the productivity and influence of the ChatGPT area and identify the major research themes in this field. Therefore, we posed the following research questions (RQs):
  • RQ1: Who/what are the most productive researchers, institutions, and countries/territories in the ChatGPT field?
  • RQ2: What is the status of academic collaborations among researchers or countries/regions in the ChatGPT field?
  • RQ3: Which are the most impactable researchers, journals, and articles in the ChatGPT field?
  • RQ4: What are the main themes in the ChatGPT field?

2. What Is Bibliometric Analysis?

Bibliometric analysis refers to a reliable methodology for objectively assessing the productivity and influence of a specific scientific field, and identifying primary themes in the field [13]. In general, bibliometric analysis is objective and little prone to bias because it employs automated or semi-automated quantitative data and instruments [12,14].
Due to the strengths of bibliometric analysis, several researchers have performed bibliometric analyses to objectively assess the overview or intellectual structure of particular fields [12,15,16]. More specifically, these studies applied the CiteSpace, a literature information visualization instrument, to conduct several types of bibliometric analysis such as co-citation analysis. Thus, our study also employed CiteSpace to perform bibliometric analyses of the ChatGPT area.

3. Methodologies

As reported by prior studies [12,17,18,19], the Web of Science (WoS) is a reliable bibliometric indexing platform for academic literature and has been widely used to gather academic articles in several bibliometric studies [15,20,21]. Therefore, we used WoS to collect papers related to ChatGPT.
Based on the PRISMA Flow Diagram (Figure 1), we collected the ChatGPT-related articles as follows. First, we employed the Social Science Citation Index (SSCI), Science Citation Index Expanded (SCIE), Emerging Sources Citation Index (ESCI), Arts and Humanities Citation Index (A&HCI), Conference Proceedings Citation Index–Social Science and Humanities (CPCI-SSH), and Conference Proceedings Citation Index–Science (CPCI-S), offered by the WoS Core Collection (WoSCC) database, as the primary data sources, and searched 3681 documents that included the term “ChatGPT” in their titles, abstracts, or author keywords. Then, we excluded documents such as “Editorial Materials”, “Letter”, etc., as well as documents not written in English. Consequently, 2465 ChatGPT articles were collected (retrieved on 24 March 2024).
We employed CiteSpace (Advanced Version 6.3. R1) for the data analysis. CiteSpace is one of the most popular information visualization applications [12,16] and is designed to visually explore scientific literature from a specific field [17]. The parameters listed in Table 1 were used for the analysis.

4. Results

4.1. Major Research Forces

Following the suggestions of Zhao et al. [12], we first identified the primary research forces in the ChatGPT field by examining the most productive researchers, institutions, and countries/territories. Also, to understand academic collaboration status in the field, we conducted co-author network analysis at the levels of author and country/region.

4.1.1. Researchers with High Publication Output and Author-Based Co-Authorship Network Analysis

Table 2 lists the top eleven authors/researchers according to the number of published articles, and Ishith Seth and Himel Mondal each authored thirteen papers. Currently, Ishith Seth is affiliated with Monash University and Peninsula Health, Australia. His/her thirteen articles primarily focus on investigating ChatGPT in the field of surgery [22,23]. In terms of Himel Mondal, currently affiliated with All India Institute of Medical Sciences, India, and his/her thirteen articles mainly discussed the application of ChatGPT in the context of education and academic writing [24,25]. Followed by them, Melissa A. Kacena and Fei-Yue Wang have authored twelve and eleven papers, respectively.
In addition, we conducted author-based co-authorship analysis and generated the collaboration networks as shown in Figure 2. The nodes indicate authors/researchers, while the links among nodes represent the collaborations among authors/researchers. It is worth noting that some collaborations may be between supervisors and graduate students. As shown in Figure 2, the overall network consists of several isolated sub-networks, which means that the authors work in small groups and lack communication [20]. Also, we found a large sub-network led by Melissa A. Kacena, Jill C. Fehrenbacher, and Lilian I. Plotkin.

4.1.2. Productive Institutions

Table 3 lists the top ten institutes according to the number of published papers. Stanford University is ranked first, having published 43 articles, followed by the Harvard University (38), National University of Singapore (37), and Monash University (27).

4.1.3. Productive Countries/Regions and Country/Region-Based Co-Authorship Network Analysis

Table 4 lists the top ten countries/territories according to the number of published papers. Among them, the United States has been the most productive, contributing 847 articles, followed by the People’s Republic of China (322), United Kingdom (206), and India (167).
We also performed country/region-based co-authorship network analysis (based on pathfinder network) and generated the collaboration networks as shown in Figure 3. The layout is organized so that countries/regions with strong collaborative relationships will appear near to each other, whereas countries/regions with weak collaboration ties would be placed further away from one another. The nodes indicate countries/territories, while the links among nodes represent the collaborations among countries/territories. The node size revealed the number of publications of the corresponding countries/territories. The pink node edge means that the node has strong betweenness centrality. Betweenness centrality can be illustrated as the number of times a node lies on the shortest path between others [26]. Concretely, the main findings are as follows. First, the United Kingdom, India, and Spain had the highest degree of betweenness centrality, which means that they play the most significant roles in the overall collaboration network (see Table 5). Second, although Qatar, Argentina, and Nigeria have high betweenness centrality, their number of ChatGPT-related publications is relatively low. Third, in international politics, the countries/regions with a high degree of betweenness centrality may serve as bridges between different political groups, economic alliances, or cultural districts. For example, Qatar is an economic and cultural hub connecting the West and the Middle East.

4.2. Co-Citation Analysis

As suggested by several researchers [15,27], we identified the impactful authors/researchers, journals/sources, and references/literature related to ChatGPT through co-citation analysis.

4.2.1. Highly Cited Authors/Researchers

Table 6 lists the top ten researchers/authors and their citation counts. We found that Tiffany H. Kung and Tom Brown were cited by 317 and 272 counts, respectively. Currently, Tiffany H. Kung is affiliated with AnsibleHealth, USA, and her/his research, Kung et al. [28], was also the highly cited literature in our reference co-citation analysis. Regarding Tom Brown, currently affiliated with Anthropic, USA, his/her research [29] developed GPT-3, which is the basis of ChatGPT.

4.2.2. Highly Cited Journals/Sources

Table 7 lists the top ten journals/sources and their citation counts. It is found that the ARXIV had the highest citation counts (1207) in the ChatGPT field. This source was followed by the Nature (710), Cureus Journal of Medical Science (394), and Science (375).

4.2.3. Highly Cited References/Literature

Table 8 lists the top eight references and their citation counts. Among them, Kung et al. [28] was one of the most impactful studies on ChatGPT.
A review of the impactful articles on ChatGPT (Table 5) resulted in the following key findings:
Kung et al. [28], Gilson et al. [31], and Sallam [30] investigated ChatGPT in the context of medical and healthcare education. Kung et al. [28] and Gilson et al. [31] tested the performance of ChatGPT on the United States Medical Licensing Exam, wherein it achieved scores near the passing threshold. Based on this finding, they reported that LLMs, such as ChatGPT, can potentially support medical education.
Van Dis et al. [5] and Dwivedi et al. [32] discussed the ethical usage of ChatGPT. For example, Dwivedi et al. [28] proposed that relevant agencies should promote the establishment of penalty standards for the intentional misuse and abuse of ChatGPT.
Kasneci et al. [33] and Rudolph et al. [34] focused on the integration of ChatGPT in educational contexts, albeit from different perspectives. Kasneci et al. [33] explored the overarching opportunities and challenges presented by LLMs, such as ChatGPT, in education, highlighting their potential to augment learning and teaching while cautioning against issues such as bias and over-reliance. In contrast, Rudolph et al. [34] focused on the specific impacts of ChatGPT on assessments in higher education, discussing its transformative potential as well as concerns regarding academic integrity and traditional evaluation methods. Both studies underscored the necessity of careful and responsible incorporation of AI tools in education.

4.3. Cluster Analysis

To confirm the primary themes in the field of ChatGPT, we conducted a keyword-based cluster analysis (selection criteria: g-index (k = 50), LRF = 3.0, L/N = 10, LBY = 5, e = 1.0), wherein CiteSpace generated 10 clusters (Figure 4) and labeled them using the log-likelihood ratio (LLR) algorithm [17,35] as it can generate clusters with high intraclass variability and low interclass similarity [36].
On the basis of the guidelines offered by Zhao et al. [12], we found Clusters #0 (processing artificial intelligence; e.g., Aiumtrakul et al. [37]), Cluster #1 (behavioral intention; e.g., Ma and Huo [38]; Kopplin et al. [39]), Cluster #2 (differential diagnosis list; e.g., Abdullahi et al. [40]; Hirosawa et al. [41]), to be the most significant in the ChatGPT field (Table 9) owing to their high silhouette values of >0.8, meaning excellent fit [35,42]. Additionally, these clusters were the largest among the 10 clusters [12]. Overall, “data processing using ChatGPT” (Cluster #0), “exploring user behavioral intention of ChatGPT” (Cluster #1), “applying ChatGPT for differential diagnosis” (Cluster #2) were found to be significant themes. Ma and Huo [38] analyzed survey-based data (n = 500) and found that novelty value and perceived humanness play notable roles in influencing intention to adopt ChatGPT. Hirosawa et al. [41] concludes the potential diagnostic accuracy of differential diagnosis lists generated by ChatGPT-3.5 and ChatGPT-4 for complex clinical vignettes derived from case reports in the department of general internal medicine.

4.4. Additional Analyses

The top ten research areas categorized by the WoS according to the number of articles are shown in Table 10. The outcomes indicate that the fields of Computer Science comprise 523 published articles, followed by Education Educational Research (428), General Internal Medicine (245), and Engineering (193). This indirectly indicates that ChatGPT has gained considerable interest in educational fields, that is, it has been widely applied in these fields.
Chen and Leydesdorff [44] reported that dual-map overlays can be used to easily determine whether a set of articles has integrated previous works across multiple subjects by tracing citation arcs from the concentrated landing zones of the origin branches. Based on the suggestions made by Zhang et al. [15], we generated journal dual-map overlays using CiteSpace (Figure 5) to explore the distribution of ChatGPT-related research knowledge bases. As shown in Figure 5, each node indicates the research subject categorized by CiteSpace based on WoS classifications. The different colors in the map indicate various areas, and the width of the links means the strength of the citation relationship between these fields. Specifically, the left- and right-hand sides show maps of citing and cited journals, respectively. The colorful links between the two maps indicate the cited/citing associations among journals/disciplines. Overall, the dual-map overlay shows that various disciplines have frequently referred to each other. Interestingly, it can be observed that the “Medicine, Medical, Clinical” and “Molecular, Biology, Immunology” disciplines have a strong tendency to refer to “Psychology, Education, Social”. These findings indirectly indicate that ChatGPT research is multidisciplinary.

5. Conclusions

5.1. Main Findings

This was one of the first studies to investigate the academic structure of ChatGPT research through a bibliometric approach. Its main findings are as follows:
The most productive researcher, institution, and country are Ishith Seth/Himel Mondal, Stanford University, and the United States, respectively. The highly cited researchers in this field are Tiffany H. Kung, Tom Brown, and Malik Sallam, whereas impactable sources/journals in this area are ARXIV, Nature, and Cureus Journal of Medical Science.
The reference co-citation analysis indicated that the most impactful article was published by Kung et al. [28], who proposed that LLMs, such as ChatGPT, can potentially support medical education. Because the topic discussed in the most cited literature can be considered as a significant research topic in a scientific area [12,45], we propose that investigating the role of ChatGPT in medical education is one of the most important research topics.
Based on the outcomes of author-based co-authorship analysis, we found that the overall collaboration network consists of several isolated sub-networks. It indirectly means that the researchers in this field tend to work in small groups and lack communication [20]. Intellectual property issues and restrictions on sharing research data may limit collaborations between the research groups [46]. Also, we found a large collaboration group led by Melissa A Kacena, Jill C Fehrenbacher, and Lilian I Plotkin. The country/region-based collaboration network indicates that the United Kingdom, India, and Spain had the highest degree of betweenness centrality, which means that they play the most significant roles in the overall collaboration network at the country/region level.
Moreover, the results of the cluster analysis indicated that “data processing using ChatGPT”, “exploring user behavioral intention of ChatGPT”, and “applying ChatGPT for differential diagnosis” are significant themes. We also determined that ChatGPT has been widely applied in the educational context. This is supported by the results of additional analyses (Section 4.4). It implies that ChatGPT research is multidisciplinary.
Furthermore, by performing co-authorship and co-citation analyses, our study addresses the limitations of the prior ChatGPT-related bibliometric study by Baber et al. [47], which lacked such analyses. These analyses offer deeper insights into collaborative networks and intellectual structures.
Finally, our findings offer insights for companies and investors toward developing more effective strategies for ChatGPT-related technologies, in collaboration with productive institutions and scholars in the ChatGPT field.

5.2. Future Directions

Based on the outcomes of cluster analysis and reference co-citation analysis, we mention some avenues for future research in the ChatGPT context. These directions enhance the potential of ChatGPT and similar generative AI technologies.
The integration of ChatGPT into medical education reveals a transformative opportunity. Future research can focus on developing adaptive learning platforms where ChatGPT serves as an intelligent tutor, offering tailored feedback to medical students and practitioners.
Understanding how users interact with ChatGPT is critical for improving its design and functionality. Future studies could explore behavioral patterns, trust, and others in generative AI products or services such as ChatGPT across diverse demographic and cultural groups.

5.3. Limitations

Although the contributions of this study are noteworthy, it has some limitations. First, although WoS is a widely employed source for bibliometric analysis, WoS has some problems, such as the underrepresentation of non-English publications and the exclusion of journals with a short publication history [48]. Second, the distribution of publications in our data shows fewer authors/researchers with multiple contributions than expected. This may be due to the scope of the dataset. Third, articles published earlier in the analyzed time period had more time to accumulate citations compared to those published later, which may lead to overrepresentation of earlier articles in the co-citation analyses.

Author Contributions

Conceptualization, X.Z. and J.H.K.; Formal analysis, D.N., X.Z. and C.C.; Funding acquisition, J.H.K.; Methodology, D.N., X.Z. and C.C.; Supervision, J.H.K.; Writing—original draft, D.N., S.S. and J.H.K.; Writing—review & editing, D.N. and K.R.L. All authors have read and agreed to the published version of the manuscript.

Funding

This paper was supported by SKKU Global Research Platform Research Fund, Sungkyunkwan University, 2022–2024.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data presented in this study are available on request from the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Tian, S.; Jin, Q.; Yeganova, L.; Lai, P.-T.; Zhu, Q.; Chen, X.; Yang, Y.; Chen, Q.; Kim, W.; Comeau, D.C.; et al. Opportunities and challenges for ChatGPT and large language models in biomedicine and health. Brief. Bioinform. 2024, 25, bbad493. [Google Scholar] [CrossRef]
  2. Gill, S.S.; Xu, M.; Patros, P.; Wu, H.; Kaur, R.; Kaur, K.; Fuller, S.; Singh, M.; Arora, P.; Parlikad, A.K.; et al. Transformative effects of ChatGPT on modern education: Emerging Era of AI Chatbots. Internet Things Cyber-Phys. Syst. 2024, 4, 19–23. [Google Scholar] [CrossRef]
  3. Ouyang, L.; Wu, J.; Jiang, X.; Almeida, D.; Wainwright, C.; Mishkin, P.; Zhang, C.; Agarwal, S.; Slama, K.; Ray, A.; et al. Training language models to follow instructions with human feedback. Adv. Neural Inf. Process. Syst. 2022, 35, 27730–27744. [Google Scholar]
  4. Jansen, B.J.; Jung, S.G.; Salminen, J. Employing large language models in survey research. Nat. Lang. Process. J. 2023, 4, 100020. [Google Scholar] [CrossRef]
  5. Van Dis, E.A.; Bollen, J.; Zuidema, W.; van Rooij, R.; Bockting, C.L. ChatGPT: Five priorities for research. Nature 2023, 614, 224–226. [Google Scholar] [CrossRef]
  6. Haleem, A.; Javaid, M.; Singh, R.P. An era of ChatGPT as a significant futuristic support tool: A study on features, abilities, and challenges. BenchCouncil Trans. Benchmarks Stand. Eval. 2022, 2, 100089. [Google Scholar] [CrossRef]
  7. Achiam, J.; Adler, S.; Agarwal, S.; Ahmad, L.; Akkaya, I.; Aleman, F.L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al. Gpt-4 technical report. arXiv 2023, arXiv:2303.08774. [Google Scholar]
  8. Nan, D.; Sun, S.; Zhang, S.; Zhao, X.; Kim, J.H. Analyzing behavioral intentions toward Generative Artificial Intelligence: The case of ChatGPT. Univers. Access Inf. Soc. 2024, 1–11. [Google Scholar] [CrossRef]
  9. Argyle, L.P.; Busby, E.C.; Fulda, N.; Gubler, J.R.; Rytting, C.; Wingate, D. Out of one, many: Using language models to simulate human samples. Political Anal. 2023, 31, 337–351. [Google Scholar] [CrossRef]
  10. Rice, S.; Crouse, S.R.; Winter, S.R.; Rice, C. The advantages and limitations of using ChatGPT to enhance technological research. Technol. Soc. 2024, 76, 102426. [Google Scholar] [CrossRef]
  11. Hämäläinen, P.; Tavast, M.; Kunnari, A. Evaluating Large Language Models in Generating Synthetic Hci Research Data: A Case Study. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, Hamburg Germany, 23–28 April 2023; pp. 1–19. [Google Scholar]
  12. Zhao, X.; Nan, D.; Chen, C.; Zhang, S.; Che, S.; Kim, J.H. Bibliometric study for environmental, social, and governance research using CiteSpace. Front. Environ. Sci. 2023, 10, 2534. [Google Scholar] [CrossRef]
  13. Lim, W.M.; Kumar, S.; Ali, F. Advancing knowledge through literature reviews: ‘what’, ‘why’, and ‘how to contribute’. Serv. Ind. J. 2022, 42, 481–513. [Google Scholar] [CrossRef]
  14. Donthu, N.; Kumar, S.; Mukherjee, D.; Pandey, N.; Lim, W.M. How to conduct a bibliometric analysis: An overview and guidelines. J. Bus. Res. 2021, 133, 285–296. [Google Scholar] [CrossRef]
  15. Zhang, S.; Che, S.; Nan, D.; Kim, J.H. MOOCs as a research agenda: Changes over time. Int. Rev. Res. Open Distrib. Learn. 2022, 23, 193–210. [Google Scholar] [CrossRef]
  16. Azam, A.; Ahmed, A.; Kamran, M.S.; Hai, L.; Zhang, Z.; Ali, A. Knowledge structuring for enhancing mechanical energy harvesting (MEH): An in-depth review from 2000 to 2020 using CiteSpace. Renew. Sustain. Energy Rev. 2021, 150, 111460. [Google Scholar] [CrossRef]
  17. Chen, C. CiteSpace II: Detecting and visualizing emerging trends and transient patterns in scientific literature. J. Am. Soc. Inf. Sci. Technol. 2006, 57, 359–377. [Google Scholar] [CrossRef]
  18. Sabah, N.U.; Khan, M.M.; Talib, R.; Anwar, M.; Arshad Malik, M.S.; Ellyza Nohuddin, P.N. Google Scholar University Ranking Algorithm to Evaluate the Quality of Institutional Research. Computers. Mater. Contin. 2023, 75, 4955–4972. [Google Scholar] [CrossRef]
  19. Naseer, I.; Masood, T.; Akram, S.; Jaffar, A.; Rashid, M.; Iqbal, M.A. Lung Cancer Detection Using Modified AlexNet Architecture and Support Vector Machine. Computers. Mater. Contin. 2023, 74, 032927. [Google Scholar] [CrossRef]
  20. Kim, J.H.; Nan, D. Machine learning and deep learning for social science: A bibliometric approach. In Research Handbook on Artificial Intelligence and Communication; Edward Elgar Publishing: Cheltenham, UK, 2023; p. 214. [Google Scholar]
  21. Sun, S.; Nan, D.; Che, S.; Kim, J.H. Investigating the knowledge structure of research on massively multiplayer online role-playing games: A bibliometric analysis. Data Inf. Manag. 2022, 8, 100024. [Google Scholar] [CrossRef]
  22. Seth, I.; Cox, A.; Xie, Y.; Bulloch, G.; Hunter-Smith, D.J.; Rozen, W.M.; Ross, R.J. Evaluating chatbot efficacy for answering frequently asked questions in plastic surgery: A ChatGPT case study focused on breast augmentation. Aesthetic Surg. J. 2023, 43, 1126–1135. [Google Scholar] [CrossRef] [PubMed]
  23. Xie, Y.; Seth, I.; Hunter-Smith, D.J.; Rozen, W.M.; Ross, R.; Lee, M. Aesthetic surgery advice and counseling from artificial intelligence: A rhinoplasty consultation with ChatGPT. Aesthetic Plast. Surg. 2023, 47, 1985–1993. [Google Scholar] [CrossRef]
  24. Mondal, H.; Mondal, S.; Podder, I. Using ChatGPT for writing articles for patients’ education for dermatological diseases: A pilot study. Indian Dermatol. Online J. 2023, 14, 482–486. [Google Scholar] [CrossRef] [PubMed]
  25. Mondal, H.; Mondal, S. ChatGPT in academic writing: Maximizing its benefits and minimizing the risks. Indian J. Ophthalmol. 2023, 71, 3600–3606. [Google Scholar] [CrossRef]
  26. Newman, M.E. A measure of betweenness centrality based on random walks. Soc. Netw. 2005, 27, 39–54. [Google Scholar] [CrossRef]
  27. Sabé, M.; Sulstarova, A.; Chen, C.; Hyde, J.; Poulet, E.; Aleman, A.; Downar, J.; Brandt, V.; Mallet, L.; Sentissi, O.; et al. A century of research on neuromodulation interventions: A scientometric analysis of trends and knowledge maps. Neurosci. Biobehav. Rev. 2023, 152, 105300. [Google Scholar] [CrossRef] [PubMed]
  28. Kung, T.H.; Cheatham, M.; Medenilla, A.; Sillos, C.; De Leon, L.; Elepaño, C.; Madriaga, M.; Aggabao, R.; Diaz-Candido, G.; Maningo, J.; et al. Performance of ChatGPT on USMLE: Potential for AI-assisted medical education using large language models. PLoS Digit. Health 2023, 2, e0000198. [Google Scholar] [CrossRef] [PubMed]
  29. Brown, T.; Mann, B.; Ryder, N.; Subbiah, M.; Kaplan, J.D.; Dhariwal, P.; Neelakantan, A.; Shyam, P.; Sastry, G.; Askell, A.; et al. Language models are few-shot learners. Adv. Neural Inf. Process. Syst. 2020, 33, 1877–1901. [Google Scholar]
  30. Sallam, M. ChatGPT Utility in Healthcare Education, Research, and Practice: Systematic Review on the Promising Perspectives and Valid Concerns. Healthcare 2023, 11, 887. [Google Scholar] [CrossRef] [PubMed]
  31. Gilson, A.; Safranek, C.W.; Huang, T.; Socrates, V.; Chi, L.; Taylor, R.A.; Chartash, D. How does ChatGPT perform on the United States medical licensing examination? The implications of large language models for medical education and knowledge assessment. JMIR Med. Educ. 2023, 9, e45312. [Google Scholar] [CrossRef]
  32. Dwivedi, Y.K.; Kshetri, N.; Hughes, L.; Slade, E.L.; Jeyaraj, A.; Kar, A.K.; Baabdullah, A.M.; Koohang, A.; Raghavan, V.; Ahuja, M.; et al. “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. Int. J. Inf. Manag. 2023, 71, 102642. [Google Scholar] [CrossRef]
  33. Kasneci, E.; Seßler, K.; Küchemann, S.; Bannert, M.; Dementieva, D.; Fischer, F.; Gasser, U.; Groh, G.; Günnemann, S.; Hüllermeier, E.; et al. ChatGPT for good? On opportunities and challenges of large language models for education. Learn. Individ. Differ. 2023, 103, 102274. [Google Scholar] [CrossRef]
  34. Rudolph, J.; Tan, S.; Tan, S. ChatGPT: Bullshit spewer or the end of traditional assessments in higher education? J. Appl. Learn. Teach. 2023, 6, 342–363. [Google Scholar]
  35. Pestana, M.H.; Wang, W.C.; Parreira, A. Senior tourism–A scientometric review (1998–2017). Tour. Rev. 2020, 75, 699–715. [Google Scholar] [CrossRef]
  36. Chopra, M.; Saini, N.; Kumar, S.; Varma, A.; Mangla, S.K.; Lim, W.M. Past, present, and future of knowledge management for business sustainability. J. Clean. Prod. 2021, 328, 129592. [Google Scholar] [CrossRef]
  37. Aiumtrakul, N.; Thongprayoon, C.; Arayangkool, C.; Vo, K.B.; Wannaphut, C.; Suppadungsuk, S.; Krisanapan, P.; Garcia Valencia, O.A.; Qureshi, F.; Miao, J.; et al. Personalized Medicine in Urolithiasis: AI Chatbot-Assisted Dietary Management of Oxalate for Kidney Stone Prevention. J. Pers. Med. 2024, 14, 107. [Google Scholar] [CrossRef] [PubMed]
  38. Ma, X.; Huo, Y. Are users willing to embrace ChatGPT? Exploring the factors on the acceptance of chatbots from the perspective of AIDUA framework. Technol. Soc. 2023, 75, 102362. [Google Scholar] [CrossRef]
  39. Kopplin, C.S. Chatbots in the workplace: A technology acceptance study applying uses and gratifications in coworking spaces. J. Organ. Comput. Electron. Commer. 2022, 32, 232–257. [Google Scholar] [CrossRef]
  40. Abdullahi, T.; Singh, R.; Eickhoff, C. Learning to Make Rare and Complex Diagnoses With Generative AI Assistance: Qualitative Study of Popular Large Language Models. JMIR Med. Educ. 2024, 10, e51391. [Google Scholar] [CrossRef] [PubMed]
  41. Hirosawa, T.; Kawamura, R.; Harada, Y.; Mizuta, K.; Tokumasu, K.; Kaji, Y.; Suzuki, T.; Shimizu, T. ChatGPT-generated differential diagnosis lists for complex case–derived clinical vignettes: Diagnostic accuracy evaluation. JMIR Med. Inform. 2023, 11, e48808. [Google Scholar] [CrossRef] [PubMed]
  42. Chen, C.; Dubin, R.; Kim, M.C. Emerging trends and new developments in regenerative medicine: A scientometric update (2000–2014). Expert Opin. Biol. Ther. 2014, 14, 1295–1317. [Google Scholar] [CrossRef]
  43. Widziewicz-Rzońca, K.; Tytła, M. First systematic review on PM-bound water: Exploring the existing knowledge domain using the CiteSpace software. Scientometrics 2020, 124, 1945–2008. [Google Scholar] [CrossRef]
  44. Chen, C.; Leydesdorff, L. Patterns of connections and movements in dual-map overlays: A new method of publication portfolio analysis. J. Assoc. Inf. Sci. Technol. 2014, 65, 334–351. [Google Scholar] [CrossRef]
  45. Abbas, N.N.; Ahmed, T.; Shah, S.H.U.; Omar, M.; Park, H.W. Investigating the applications of artificial intelligence in cyber security. Scientometrics 2019, 121, 1189–1211. [Google Scholar] [CrossRef]
  46. Borgman, C.L. The conundrum of sharing research data. J. Am. Soc. Inf. Sci. Technol. 2012, 63, 1059–1078. [Google Scholar] [CrossRef]
  47. Baber, H.; Nair, K.; Gupta, R.; Gurjar, K. The beginning of ChatGPT–a systematic and bibliometric review of the literature. Inf. Learn. Sci. 2024, 125, 587–614. [Google Scholar] [CrossRef]
  48. Mongeon, P.; Paul-Hus, A. The journal coverage of Web of Science and Scopus: A comparative analysis. Scientometrics 2016, 106, 213–228. [Google Scholar] [CrossRef]
Information 16 00038 g001
Figure 1. PRISMA Flow Diagram.
Figure 1. PRISMA Flow Diagram.
Information 16 00038 g001
Information 16 00038 g002
Figure 2. Author-based collaboration network.
Figure 2. Author-based collaboration network.
Information 16 00038 g002
Information 16 00038 g003
Figure 3. Country/region-based collaboration network. Note: Taiwan is considered a region.
Figure 3. Country/region-based collaboration network. Note: Taiwan is considered a region.
Information 16 00038 g003
Information 16 00038 g004
Figure 4. Cluster analysis.
Figure 4. Cluster analysis.
Information 16 00038 g004
Information 16 00038 g005
Figure 5. Dual-map overlay of ChatGPT research on the Journal Citation Reports map.
Figure 5. Dual-map overlay of ChatGPT research on the Journal Citation Reports map.
Information 16 00038 g005
Table 1. Parameter settings.
Table 1. Parameter settings.
Parameters
Time slicingMonthly (y = 0)/Yearly (y = 1)
PruningPathfinder & Pruning sliced networks
Node typeauthor, institution, country, reference, cited author, cited journal, keyword
LinksStrength (cosine) & Scope (within slices)
Selection criteriag-index (k = 25)/g-index (k = 50)
VisualizationStatic cluster view & Merged network
Note: The g-index aims to give more weight to articles with high citation counts.
Table 2. Top eleven authors according to the number of published ChatGPT-related papers.
Table 2. Top eleven authors according to the number of published ChatGPT-related papers.
Number of PapersAuthor/Researcher
13Seth, Ishith
13Mondal, Himel
12Kacena, Melissa A
11Wang, Fei-Yue
9Fehrenbacher, Jill C
9Rozen, Warren M
8Cheungpasitporn, Wisit
7Klang, Eyal
7Mondal, Shaikat
7Sohail, Shahab Saquib
7Elyoseph, Zohar
Table 3. Top ten institutes according to the number of published ChatGPT-related articles.
Table 3. Top ten institutes according to the number of published ChatGPT-related articles.
Number of PapersInstitution
43Stanford University
38Harvard University
37National University of Singapore
27Monash University
25Indiana University Bloomington
24Chinese Academy of Sciences
22University of Toronto
20University of Hong Kong
20Imperial College London
20University of Pennsylvania
Table 4. Top ten countries/territories according to the number of ChatGPT-related publications.
Table 4. Top ten countries/territories according to the number of ChatGPT-related publications.
Number of PapersCountry/TerritoryBetweenness Centrality
847United States0.02
322People’s Republic of China0.01
206United Kingdom0.15
167India0.14
153Germany0.05
141Australia0.03
103Italy0.07
100Canada0.07
89Spain0.12
88Turkiye0.03
Table 5. Top eight countries/territories according to the degree of betweenness centrality.
Table 5. Top eight countries/territories according to the degree of betweenness centrality.
Betweenness CentralityCountry/TerritoryNumber of Papers
0.15United Kingdom206
0.14India167
0.12Spain89
0.12United Arab Emirates66
0.10France51
0.10Qatar24
0.09Brazil45
0.09Argentina3
0.09Nigeria13
Table 6. Top ten authors/researchers according to the number of citations.
Table 6. Top ten authors/researchers according to the number of citations.
Number of CitationsResearcher
317Kung, Tiffany H
272Brown, Tom
271Sallam, Malik
251Vaswani, Ashish
229Gilson, Aidan
219Bockting, Claudi L
212Devlin, Jacob
204Stokel-Walker, Chris
203Dwivedi, Yogesh K
181Thorp, H Holden
Table 7. Top ten journals/sources according to the number of citations.
Table 7. Top ten journals/sources according to the number of citations.
Number of CitationsJournals/Sources
1207ARXIV
710Nature
394Cureus Journal of Medical Science
375Science
362JMIR Medical Education
346PLOS Digital Health
315Radiology
314Healthcare-Basel
311Journal of Medical Internet Research
294MEDRXIV
Table 8. Top eight ChatGPT-related articles according to the number of citations.
Table 8. Top eight ChatGPT-related articles according to the number of citations.
Number of CitationsArticle Journal/Conference Proceedings
315Kung et al. [28]PLoS digital health
231Sallam [30] Healthcare
229Gilson et al. [31] JMIR Medical Education
218Van Dis et al. [5]Nature
180Dwivedi et al. [32]International Journal of Information Management
162Kasneci et al. [33]Learning and Individual Differences
149Rudolph et al. [34] Journal of Applied Learning and Teaching
148Brown et al. [29]Advances in neural information processing systems/34th Conference on Neural Information Processing Systems (NeurIPS 2020)
Table 9. Main themes in the field of ChatGPT research.
Table 9. Main themes in the field of ChatGPT research.
Cluster IDSizeSilhouette ValueLabel (LLR)
01520.874processing artificial intelligence
11290.851behavioral intention
21120.863differential diagnosis list
Note: “Size” denotes the number of keywords in a cluster and “Silhouette value” denotes cluster homogeneity [43].
Table 10. Research areas ranked by the number of published ChatGPT-related articles.
Table 10. Research areas ranked by the number of published ChatGPT-related articles.
Number of Publications Research Area
523Computer Science
428Education Educational Research
245General Internal Medicine
193Engineering
127Health Care Sciences Services
121Surgery
94Business Economics
94Science Technology Other Topics
92Social Sciences Other Topics
83Medical Informatics
Note: The research areas are categorized based on the WoS.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Nan, D.; Zhao, X.; Chen, C.; Sun, S.; Lee, K.R.; Kim, J.H. Bibliometric Analysis on ChatGPT Research with CiteSpace. Information 2025, 16, 38. https://doi.org/10.3390/info16010038

AMA Style

Nan D, Zhao X, Chen C, Sun S, Lee KR, Kim JH. Bibliometric Analysis on ChatGPT Research with CiteSpace. Information. 2025; 16(1):38. https://doi.org/10.3390/info16010038

Chicago/Turabian Style

Nan, Dongyan, Xiangying Zhao, Chaomei Chen, Seungjong Sun, Kyeo Re Lee, and Jang Hyun Kim. 2025. "Bibliometric Analysis on ChatGPT Research with CiteSpace" Information 16, no. 1: 38. https://doi.org/10.3390/info16010038

APA Style

Nan, D., Zhao, X., Chen, C., Sun, S., Lee, K. R., & Kim, J. H. (2025). Bibliometric Analysis on ChatGPT Research with CiteSpace. Information, 16(1), 38. https://doi.org/10.3390/info16010038

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop