نوع مقاله : مقاله پژوهشی
نویسندگان
1 استاد، گروه علم اطلاعات و دانش‌شناسی، دانشکده روانشناسی و علوم تربیتی، دانشگاه خوارزمی، تهران، ایران.
2 دکتری، علم اطلاعات و دانش‌شناسی، دانشکده روانشناسی و علوم تربیتی، دانشگاه خوارزمی، تهران، ایران.
3 استادیار، گروه علم اطلاعات و دانش‌شناسی، دانشکده روانشناسی و علوم تربیتی، دانشگاه خوارزمی، تهران، ایران.
چکیده
کلیدواژه‌ها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Purpose: Artificial intelligence (AI) has profoundly impacted the field of Library and Information Science (LIS) by transforming traditional systems of knowledge management, information services, and user interaction. As AI technologies evolve rapidly, from machine learning and natural language processing (NLP) to large language models (LLMs) and generative AI, understanding the intellectual landscape of AI research in LIS becomes increasingly crucial. This study aims to identify, map, and analyze the thematic and citation structures of AI-related publications in LIS to uncover dominant research clusters, emerging trends, and potential gaps.
Methodology: This applied scientometric study utilized a mixed-methods approach based on co-citation and keyword co-occurrence analysis. A dataset of 3,066 records published from 2005 to 2024 was retrieved from the Web of Science Core Collection using the query: (WC=Library and Information Science) AND (TS=AI OR Artificial Intelligence). Co-citation networks were constructed using CiteSpace software over four five-year intervals with a Top N = 50 threshold. Cluster quality was evaluated using Modularity (Q) and Silhouette (S) indices. Labels for clusters were generated through software algorithms, manual document inspection, and expert consensus involving five specialists in AI and LIS. Burst detection was applied to highlight influential articles. Keyword co-occurrence networks were generated via VOSviewer with a controlled vocabulary. Hot topic trends and keyword trajectories were visualized using Biblioshiny from the Bibliometrix R package.
Findings: The co-citation analysis revealed 24 thematic clusters, with five key clusters including algorithmic decision-making (2016), the moderating role of AI (2019), explainable AI (2019), systematic AI reviews (2020), and AI applications in public services (2020). A review of highly cited articles indicated that Dwivedi et al. (2021), whose article explored multidisciplinary perspectives on AI, had the greatest impact with 133 citations. He was followed by Duan et al. (2019), whose study focused on the role of AI in decision making, with 107 citations, and Sun & Medaglia (2019), who examined the application of AI in the public sector, with 75 citations. The burstiness index revealed that Russell & Norvig (2016), author of the influential textbook "Artificial Intelligence: A Modern Approach", showed the most significant growth in impact. The keyword co-occurrence map identified five key clusters: information systems and knowledge management, technology adoption and user interaction, machine learning and AI, language models and information services, and AI ethics and governance. These clusters indicate research trends focused on data management, user behavior analysis, model development, information service improvement, and ethical challenges. A historical trend analysis shows a shift from traditional concepts such as ontology and digital libraries to machine learning, NLP, and LLMs. Today, research primarily focuses on generative AI, transparency, and governance. Additionally, highly cited terms in this domain include Industry 4.0, decision-making, and big data analytics, highlighting AI’s role in enhancing human interactions and data management. Moreover, trend analysis results indicate that research in AI within LIS has shifted from classical concepts to advanced technologies such as machine learning and neural networks. The focus has moved from data mining and information retrieval toward deep learning and NLP. Recently, large language models and generative AI have brought fundamental transformations to this field. Additionally, ethical concerns and algorithmic transparency have gained increasing importance. This trend signifies a transition toward smarter, more automated, and language-processing-oriented systems, with growing attention to ethical considerations.
Conclusion: The findings indicate that research in AI within LIS is primarily focused on five main areas: data management, technology adoption, machine learning, language models, and ethical issues. The interconnections between these clusters suggest that the successful implementation of AI requires effective coordination between data management, user behavior analysis, machine learning model development, and ethical compliance. These insights can assist policymakers, researchers, and professionals in optimizing the use of AI technologies in information systems. Furthermore, recent developments highlight the emergence of AI literacy and university libraries associated with modern technologies. Emphasis on user acceptance, trust in AI systems, and their impact on information policies suggests that AI’s role in LIS extends beyond technical tools to encompass social, ethical, and policy dimensions. This trend illustrates a shift from classical topics toward intelligent interactions, big data analytics, and the adoption of emerging technologies, emphasizing the need for further research on AI governance, transparency, and social implications. Overall, this study helps identify strong and weak research areas in this field, highlighting the necessary capacities for enhancing research and promoting responsible AI development in LIS. Ultimately, scientometric analysis can aid policymakers and planners in optimizing resource allocation, improving the socio-economic structure, and advancing sustainable AI development.
کلیدواژه‌ها [English]
)