Publications on COVID-19 and artificial intelligence: trends and lessons
Article information
Abstract
Purpose
This study investigates shifts in scientific research focus, particularly the decline in COVID-19-related research and the rapid growth of artificial intelligence (AI) publications.
Methods
We analyzed publication data from the Web of Science, comparing yearly publication counts for COVID-19 and AI research. The study also tracked changes in the impact factors of leading journals like Science and Nature, alongside those of top AI journals over the past decade. Additionally, we reviewed the top 10 most cited articles in 2021 from Science and Nature and the most influential AI publications from the past five years according to Google Scholar. The impact trends of the top 100 AI journals in computer science were also explored.
Results
The analysis reveals a noticeable decline in COVID-19 related publications as the pandemic urgency diminishes, contrasted with the continued rapid growth of AI research. Impact factors for prestigious journals have shifted, with AI journals increasingly dominating the academic landscape. The review of top-cited articles further emphasizes these trends.
Conclusion
Our findings indicate a significant shift in research priorities, with AI emerging as a dominant field poised to address future challenges, reflecting the evolving focus of the scientific community.
Introduction
Background
The onset of the COVID-19 pandemic in late 2019 and early 2020 catalyzed an unprecedented surge in scientific research. Faced with a global health crisis of unparalleled scale, researchers worldwide mobilized rapidly to understand the novel coronavirus, SARS-CoV-2, and to develop strategies for its containment and treatment [1]. This urgent need for knowledge spurred a prolific increase in publications, as scientists and healthcare professionals raced to share findings and insights. The resulting body of literature on COVID-19 has been vast, encompassing a broad range of topics from epidemiology and virology to public health measures and vaccine development [2].
However, as the immediate threat of the pandemic has begun to wane due to successful vaccination campaigns and improved therapeutic protocols, the volume of COVID-19–related research outputs has shown a noticeable decline [3]. This trend is indicative of the natural progression of scientific focus, where acute phases of crisis are often followed by periods of consolidation and reflection. Researchers are now evaluating the long-term implications of the pandemic, studying its socioeconomic impacts, and considering how to better prepare for future pandemics [4]. Nevertheless, the peak period of COVID-19 research has passed, marking a shift in the collective scientific endeavor.
In stark contrast to this decline, the field of artificial intelligence (AI) has been experiencing sustained and rapid growth in academic publications. Over the past decade, AI has transitioned from a specialized niche to a dominant area of research with widespread applications across numerous disciplines [5]. The integration of AI technologies in healthcare, finance, transportation, and other sectors has demonstrated the transformative potential of machine learning, neural networks, and other AI methodologies. This burgeoning interest is reflected in the increasing number of AI-related publications, which continue to rise sharply even as COVID-19 research declines.
The expanding influence of AI in the academic landscape can be attributed to several factors. Firstly, advancements in computational power and data availability have enabled more sophisticated and scalable AI models [6]. Secondly, the versatility of AI applications—from natural language processing and computer vision to predictive analytics and autonomous systems has attracted interest from a diverse array of research fields [7]. Lastly, the potential for AI to address complex and multifaceted challenges has galvanized a broad community of researchers to explore its capabilities and implications.
Objectives
In this study, we aim to examine these publication trends in detail by analyzing data from the Web of Science. We focus on the annual publication counts for COVID-19 and AI research to illustrate the contrasting trajectories of these fields. This analysis not only highlights the dynamic nature of scientific research but also underscores the shifting priorities within the academic community. We further investigate the impact factor changes for two prestigious scientific journals, Science and Nature, which have played pivotal roles in disseminating high-impact research findings throughout the pandemic. By tracking the impact factor trends of leading AI journals over the past decade, we provide insights into the growing influence and recognition of AI research within the scientific community. To gain a deeper understanding of the specific contributions that have shaped these trends, we review the top 10 most cited articles in 2021 from Science and Nature. Additionally, we examine the most influential publications from the past 5 years according to h5-index metrics, offering a perspective on the pivotal works driving research impact in these fields. Finally, we explore the impact trends of the top 100 AI journals in computer science. This comprehensive analysis sheds light on the evolving focus of scientific research, with AI emerging as a dominant and rapidly advancing field poised to address future global challenges. By presenting these findings, we aim to provide valuable insights into the shifting landscape of scientific research and highlight the lessons learned from the COVID-19 pandemic. The growing prominence of AI research underscores the importance of continued investment in this field to harness its potential for addressing complex problems and driving innovation across various sectors.
Methods
Ethics statement
This study did not involve human subjects, and therefore ethics approval was not required.
Study design
This descriptive study was based on an online indexing database.
Data collection
The publication trend of major journals and the impact of each journal were analyzed using metrics that assess the value and influence of journals. The Web of Science Core Collection database and Journal Citation Reports (JCR) database provided by Clarivate were used [8]. Journal trends were based on the data up to June 30, 2024, and the JCR database included citation indexes from the Science Citation Index Expanded (SCIE) and Emerging Sources Citation Index (ESCI). For the analysis of AI journal impact trends, we selected major AI-related journals categorized as “artificial intelligence,” “theory & methods,” “information systems,” and “information science & library science” within the “computer science” category.
Statistical analysis
Descriptive statistics were presented to analyze the results of journal impact trends.
Results
Fig. 1 shows the number of publications per year related to COVID-19 (or SARS-CoV-2) and AI from the Web of Science database. For COVID-19, there is a sharp increase in publications starting in 2020, peaking in 2022, followed by a decline in 2023 and 2024. The surge in publications corresponds with the onset of the COVID-19 pandemic, with a peak during the height of the global health crisis. The subsequent decline reflects a decrease in new research outputs as the immediate urgency of the pandemic subsided. For AI, there is a steady increase in related publications from 2010 to 2020, followed by a sharp rise in 2021, peaking in 2023. The consistent growth in AI research indicates a sustained interest and advancements in the field. The sharp increase in recent years reflects the expanding applications and significance of AI across various disciplines. Fig. 1 reveals a stark contrast between the two fields. COVID-19-related research shows a rapid surge followed by a decline, typical of a response to an urgent global event. In contrast, AI research shows steady and accelerating growth, reflecting its long-term and broad-based impact on multiple sectors. The data indicates a shift in scientific focus, while COVID-19 research peaked and is now decreasing, AI research is on a robust upward trajectory, suggesting its increasing dominance and relevance in the scientific community. These trends provide insights into how global events and technological advancements influence research priorities and publication outputs over time.
The huge impact of COVID-19 on academic publications is evident in the change in the impact factors of the world’s leading multidisciplinary journals. Fig. 2 depicts the change in impact factors for the prestigious journals Science and Nature from 2010 to 2023, with a highlighted focus on the COVID-19 pandemic period. From 2010 to around 2018, both journals show a general upward trend in impact factors with some fluctuations. The lines for Science and Nature often move closely together, indicating similar trends in their impact factors. During the COVID-19 period, there was a noticeable spike in the impact factors of both journals. This is likely due to the high volume of critical and high-impact research published related to COVID-19, which garnered significant number of citations. Both journals reach their peak impact factors around 2021, aligning with the height of the pandemic and the intense focus on pandemic-related research. After peaking, both journals experience a decline in impact factors from 2022 onwards. This decline reflects the tapering-off of COVID-19-related publications as the immediate crisis subsided and the focus of research diversified. The highlighted period shows the significant influence of the COVID-19 pandemic on scientific publishing. The surge in high-impact publications during the pandemic substantially increased the impact factors of both Science and Nature. The impact factors are adjusting to the new norms of post–COVID-19, but the overall upward trend before the pandemic indicates strong resilience and the continued importance of these journals in publishing high-impact scientific research. This data illustrates how global events like the COVID-19 pandemic can dramatically affect the publication and citation dynamics within leading scientific journals.
Table 1 shows the list of the top 10 most cited articles in 2021 from Science and Nature, which illustrates the overwhelming focus on COVID-19 research during that period. COVID-19 was a global crisis with significant impacts on health, economies, and daily life. This unprecedented situation drove extensive research to understand the virus, its effects, and ways to combat it. Due to the urgent need for information, studies on COVID-19 were published rapidly and attracted considerable attention from researchers, policymakers, and the public. The dominance of COVID-19 in the top-cited articles of 2021 in Science and Nature reflects the pandemic’s unprecedented global impact and the urgent need for research across multiple scientific disciplines. The articles’ focus on viral structure, spread, public health responses, potential treatments, and prevention highlights the breadth and depth of scientific inquiry driven by the pandemic.
The impact of COVID-19 on scientific publications was indeed profound and unprecedented, but it was also relatively short-lived in terms of its peak. In contrast, the impact of AI on scientific research is growing rapidly and is expected to have a lasting and increasingly dominant effect across many fields. While COVID-19 will continue to be a topic of research, especially in the context of long-term effects, variants, and vaccine development, the immediate surge in publications related to the pandemic has decreased. The role of AI is expected to grow as the technology becomes more sophisticated and as its applications broaden. AI’s ability to handle large datasets, optimize complex processes, and generate insights is likely to make it a central tool in research across all scientific disciplines. Such an expanding role in research highlights a shift towards a more data-driven and computational approach to scientific discovery, promising to reshape the landscape of research in the coming years.
Table 2 lists the top-cited journals over the last 5 years according to Google Scholar. The h5-index is a metric used by Google Scholar to measure the impact and productivity of academic journals or conferences. It counts the number of articles published in the last 5 years that have at least that many citations. For example, an h5-index of 440 for the Institute of Electrical and Electronics Engineers (IEEE)/Computer Vision Foundation (CVF) Conference on Computer Vision and Pattern Recognition means that there are at least 440 articles from this conference published in the last 5 years that each have at least 440 citations. The high h5-index values for these AI-related conferences suggest that they are producing highly cited research. This indicates a significant and growing influence of AI research, with many impactful papers being published. The data shows that AI research is not only prolific but also highly impactful, as evidenced by the high citation counts of relevant conferences. This reflects the growing presence and influence of AI in the scientific community and beyond.
Fig. 3 exhibits the change in impact factors of major AI journals over the past decade. The impact factors of four major AI journals in AI-related categories in computer science have all increased significantly over the past decade. In 2014, the journal impact factors on average were around 5, but they exceeded 20 in 2023. The trends show the expanding interest and advancements in the field of AI and machine learning, particularly in the application area of data and signal analysis. Following this trend, IEEE Communications Surveys and Tutorials is leading in terms of citation impact. The rising impact factors of these journals indicate a thriving and evolving field with significant academic and practical impact. It reflects the increasing relevance of the research topics covered, the growing influence of the journals, and the expanding scope of interdisciplinary collaboration and innovation. For the academic community, this means a higher standard of research dissemination, greater opportunities for researchers, and an ongoing evolution of the academic landscape in these fields.
More detailed data can be seen in Table 3. It shows trends in the influence of the top 100 journals in computer science with a focus on AI. The table presents various metrics for a journal or a set of journals from 2017 to 2023, showcasing their impact and influence over time. Metrics include total citations, journal impact factor (JIF), 5-year JIF, JIF without self-citations, immediacy index, article influence score, and Eigenfactor score [9]. The ratio column indicates the relative change from 2017 to 2023. The significant increases in the metrics, particularly in total citations, JIF, 5-year JIF, and article influence score, suggest that the field of AI is experiencing rapid growth and gaining substantial recognition. The table highlights the growing impact and influence of AI research, as evidenced by the increasing citation metrics and influence scores over the years. This trend underscores the expanding importance and recognition of AI in the scientific and broader research community.
Discussion
The surge in COVID-19 research publications occurred as a direct response to the immediate and unprecedented global health crisis. As the pandemic’s acute phase evolved, and more solutions, treatments, and vaccines became available, the intensity of new research declined. The rapid rise in publications was driven by the urgent need for information on the virus, treatment options, public health strategies, and vaccine development. As the pandemic situation stabilized with the development of vaccines and effective treatments, the focus shifted from urgent research to ongoing monitoring and longitudinal studies. This led to a decrease in the volume of new research articles. In contrast, AI research has seen a continuous upward trajectory in publications, reflecting the growing importance and breadth of the field. AI’s expansion is driven by its broad applicability across various sectors, including healthcare, finance, transportation, and more. Continuous advancements in algorithms, machine learning techniques, and real-world applications fuel the ongoing increase in research output. The COVID-19 research surge was a response to a specific, high-priority global crisis with a finite timeframe, leading to an initial peak followed by a decline as the crisis was managed. AI research, however, is part of a broader and ongoing technological evolution with sustained growth and increasing relevance across many disciplines.
Research on COVID-19 had a direct and immediate impact on public health policies, treatment strategies, and emergency responses. The focus was on addressing an urgent global issue with tangible, short-term outcomes. AI research contributes to long-term technological advancements. The steady increase in publications reflects the ongoing relevance and expanding applications of AI technology. The contrasting trends illustrate how research in different fields can have varying types of impacts. Immediate crises demand rapid responses, while technological advancements like AI build over time, affecting multiple domains and creating long-term benefits. More research on COVID-19 and AI are focused on how AI technology can be used to find variables related to clinical risk, so as to provide more effective treatment methods and diagnostic parameters for infectious disease study [10]. On the one hand, realtime contact tracing applications are one of the many AI applications being used to control the virus’s spread and bolster the public health response [11]. On the other hand, AI technology could find and calculate complex patterns in large data sets on epidemics. AI tools implemented in healthcare systems for clinical practice and medical research have proved their effect for a precise identification of predictive performance models for diseases’ severity and mortality [12]. These data could inform the ongoing updating of clinical practice guidelines and facilitate physician decision-making, such as the development of targeted therapies and therapies [13]. For example, AI-assisted pneumonia products are trained on the basis of the pneumonia data and are used to assist in the diagnosis and quantitative analysis of COVID-19 pneumonia, and their clinical value and application scenarios in the post-epidemic era have been explored [14]. However, it was also found that AI has certain limitations in the application of infectious diseases. For instance, some algorithms cannot guarantee the optimal stratification or global classification performance of risks with low accuracy or conflicting results. In this sense, the future research focus is to emphasize the validation of models through external and large data sets must be mandatory for the widespread adoption of clinical predictions [15].
In summary, the stark contrast between the publication trends in COVID-19 and AI reflects differences in research urgency, focus, funding, and impact. COVID-19 research experienced a peak and subsequent decline due to the nature of the global health crisis, while AI research has shown continuous growth driven by ongoing technological developments and applications. Emerging areas like battery technology [16], carbon neutralization [17], neuromorphic computing [18], quantum computing [19], and advanced materials [20] are at the forefront of scientific research and technological innovation. Each field has unique implications for sustainability, computational power, and economic growth, illustrating the dynamic and evolving nature of research and its impact on society. Understanding these trends provides insights into how different types of research respond to immediate needs versus long-term innovation and the impact of funding and resource allocation on research dynamics.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Funding
This work was supported by the Nano and Material Technology Development Program (No. RS-2024-00405016) through the National Research Foundation of Korea (NRF), funded by the Korean Ministry of Science and ICT.
Data Availability
Most of the raw data in this paper are various indicators of Journal Citation Reports and the Web of Science database in Clarivate. Please contact the corresponding author for data availability, because Clarivate does not allow researchers to share the data.
Supplementary Materials
The authors did not provide any supplementary materials for this article.