Positive and Negative Topics of Metaverse Library from China: A Sentiment-Topic Analysis

2.1 Metaverse Libraries

The metaverse library is a new concept that emerged during the transformation of smart libraries, leveraging metaverse-related technologies for application in libraries. It encompasses a new generation of libraries that provide intelligent and personalized services to users, relying on information technologies such as artificial intelligence, the Internet of Things, virtual reality, augmented reality, 3D technology, and Edge Computing (Huynh-The et al. 2023; Baidya and Moh 2024). Research on the metaverse libraries originates from early explorations of library establishment in Second Life. At that time, Second Life was considered a precursor to the metaverse (Gent 2022), and the libraries established within it attracted the attention of scholars (Boulos et al. 2007; Feng et al. 2009). The metaverse library during this period is a “library in the metaverse,” parallel to the real world. With VR headsets, haptic gloves, AR, and extended reality (XR), the metaverse gradually evolves into an interlayer between virtual and real worlds, exhibiting characteristics of virtual-real interaction.

Scholars in the field of library science have begun to focus on the application of metaverse-related technologies in libraries (Guo et al. 2024) as well as contemplate how to transform real libraries into metaverse libraries that can provide intelligent services and virtual-real interactive experiences (Noh 2023). Some scholars have creatively proposed information retrieval algorithms for metaverse libraries, such as the Top-K Best Query Algorithm (Wu et al. 2023a) and the R-Tree-based Top-K Nearest Neighbor Book Searching Algorithm (Wu et al. 2023b). These studies provide technical support for constructing and optimizing metaverse libraries, enhancing the user experience through efficient search algorithms.

The metaverse library enables users to socialize, participate in events with authors, and receive book recommendations from AI-based systems (De Lorenzis et al. 2023). Some scholars explore relevant cases and affirm the potential of applying the metaverse concept to libraries (Noh 2023), while others conduct the user acceptance of metaverse libraries (Adetayo et al. 2023). However, overall, research on metaverse libraries remains limited. It necessitates more in-depth exploration to advance the development vision for the metaverse library.

2.2 Methods of Review Studies on Metaverse Libraries

The concept of metaverse sparks intense discussions in the library community. Numerous scholars conduct in-depth explorations of the development of metaverse from various perspectives, including development direction, construction pathways, application scenarios, and spatial construction (Yang et al. 2021; Liang et al. 2023; Zhao et al. 2022; Zhou et al. 2023).

Some scholars conduct comprehensive reviews of relevant research to summarize the academic viewpoints on metaverse libraries. Fang and Cao (2023) synthesize literature analysis, content analysis, and case studies to delineate the theoretical research and practical progress in the development of the metaverse and libraries, concluding that most current research is forward-looking, exploring the new opportunities that information technology brings to the applications of the metaverse in libraries. Yan et al. (2023) summarize existing research from theoretical and practical perspectives, reviewing metaverse library literature regarding cognitive enhancement, integration fields, application scenarios, technological foundations, and practical cases. Xing et al. (2023) conducted bibliometric and thematic analysis to summarize and compare research topics related to metaverse libraries, revealing that metaverse library-related studies in China focus on theoretical construction, while foreign studies emphasize practical exploration. From a functional positioning perspective, Li et al. (2023) conducted a literature analysis to review and summarize research findings on metaverse libraries, concluding that incorporating the metaverse leads to new changes in libraries’ preservation, educational, cultural, and recreational functions. Zhao and Lai (2023) review existing literature and find that research on metaverse libraries achieves preliminary results in theoretical research and application strategy research but also identifies issues such as a tendency towards technological pragmatism, neglect of the human element, and insufficient reflection.

These review studies systematically summarize and categorize the theoretical discussions and practical developments in this field, promoting theoretical exploration and development pathways of metaverse libraries. However, most research ignores the sentiment factors expressed in the topics of the literature. This study analyzes existing literature using a fine-grained thematic analysis method based on sentiment recognition and literary sentences.

2.3 The Sentiment Analysis of the Literature

Sentiment analysis, or opinion mining, primarily involves mining and analyzing subjective texts to extract key viewpoints and identify sentiment orientations (Zhou et al. 2020). According to principles, it can be classified into methods based on machine learning, sentiment lexicons, and hybrid methods. Among these, machine learning-based sentiment analysis searches for patterns in pre-encoded texts using machine learning techniques to classify the sentiment of uncoded texts (Dave et al. 2003).

Sentiment analysis is commonly employed in social media to analyze user comments for public opinion evolution analysis (Wang et al. 2024), market prediction (Zhang et al. 2023), and personalized recommendations (You et al. 2023). In academic literature, sentiment analysis is typically used in citation sentiment analysis, which involves analyzing and mining viewpoints, sentiments, attitudes, and sentiments from citation texts to assess the influence of cited literature (Dong and Wu 2021). Some studies directly analyze the sentiment of the content of the literature. Chen (2024) analyzes sentiment tendencies in the medical research literature and identifies four positive sentiments in the medical literature: trust, anticipation, joy, and surprise, in descending order of sentiment intensity. Moreover, Kosnik (2023) found significant differences between male and female authors in terms of the certainty-exploratory of academic perspectives and sentiment scores in discussions on present-past content. Another research study analyzed sentiment articles in the journal Science from 1997 to 2022. It reveals that academic writing has become more positive over this period and discusses its reasons from the perspectives of science popularization, publishing competition, and positive results reporting bias (Yuan and Yao 2022). Furthermore, scholars also identify positive and negative sentiment content in academic research related to the metaverse to understand this concept better (Tunca et al. 2023). These studies demonstrate the feasibility of applying sentiment analysis to research literature and suggest that identifying the sentiment content in metaverse libraries-related literature facilitates researchers in better understanding the academic perspectives on the development of metaverse libraries.

Therefore, this study adopts a machine learning-based sentiment analysis method, training models to learn how to recognize and classify sentiments from data. Compared to lexicon-based sentiment analysis methods, the former addresses issues such as limited lexicon coverage and simplistic handling of negation words and adverbs of degree, thus demonstrating better accuracy. A commonly used machine learning-based sentiment analysis method is SnowNLP, which is widely used in academic research (Ding and Li 2023; Lu et al. 2022). SnowNLP is a Python natural language processing library that can quickly and efficiently process Chinese text content and comes with pre-trained corpora. The SnowNLP library can be directly used for text sentiment classification (Zhang and Shi 2022). This study employs SnowNLP for text sentiment analysis. It is worth noting that SnowNLP training data consists of product review comments. Applying it to other fields requires fine-tuning with annotated texts from the specific field.

2.4 Topic Modeling

Topic modeling originates from information retrieval and natural language processing as a technique for data dimensionality reduction and feature extraction (Fu et al. 2019). It introduces the concept of topics by analyzing a collection of documents and identifying patterns of words and phrases, thereby categorizing the vocabulary in the document collection into topic dimensions to achieve dimensionality reduction of high-dimensional data. Meanwhile, topics contain latent semantic information about documents and their vocabulary, thus providing stronger semantic representation capabilities (Lu 2024). Common topic modeling includes Latent Semantic Analysis (LSA), Probabilistic Latent Semantic Analysis (pLSA), and Latent Dirichlet Allocation (LDA), all of which are commonly used topic modeling methods (Xia et al. 2019). In recent years, with the rapid development of deep learning algorithms and large language models, emerging topic modeling techniques such as Top2Vec and BERTopic have been widely applied in text topic mining. Among them, BERTopic uses a BERT-based deep learning pre-trained models combined with embedding models like Sentence-Transformers and the c-TF-IDF algorithm to encode and compute sentences, achieving document topic clustering and topic representation at the semantic level (Angelov 2020). Compared to traditional methods like LDA topic modeling, BERTopic performs better topic identification on the NPMI (Normalized Pointwise Mutual Information) metric (Grootendorst 2022).

3.1 Data Collection

3.2 Data Analysis

4.1 Sentiment Analysis Results

During the sentiment analysis stage, the sentiment scores for the academic text fields were calculated and statistically analyzed per document. As shown in Figure 2, the sentiment score curve is presented based on the publication dates. The collected literature spans from November 16, 2021, to January 25, 2024. Among the 91 works analyzed, 92.3 % conveyed a positive sentiment, showing strong support for metaverse libraries from the academic community in China.

Figure 2:

Literature sentiment score.

The academic community’s positive inclination towards metaverse libraries is quite evident at the literature level. Positive and negative literatures appear interspersed without clear temporal trends. It is worth noting that even though most literature expresses positive sentiments towards metaverse libraries, there are still negative segments within the overall positive sentiment literature. Therefore, further analysis at the sentence level is necessary. In this study, a total of 3,997 sentences containing both “metaverse” and “library” are extracted from the literature above. After excluding sentences with sentiment scores between 0.4 and 0.6, 2,597 positive and 1,089 negative sentiment sentences remained, subjected to topic modeling.

4.2 Topic Modeling Results

This study employs BERTopic topic modeling to extract topics from sentences expressing positive and negative sentiments, resulting in 19 positive and 9 negative sentiment topics. To improve the interpretability and accuracy of the topic descriptions, the results of the topic modeling are manually processed (Ye et al. 2023; Cui et al. 2022; Wang et al. 2024). This process involved removing invalid topics and merging similar topics. Ultimately, 10 topics are summarized for the positive and 7 for the negative dimensions.

4.2.1 Positive Dimension Topic

The manually merged topics of the positive dimensions and the corresponding thematic keywords for each topic are shown in Table 1:

Table 1:

Positive dimension topics after manual treatment.

Topic	Top 4 keywords	Keywords
Intelligent Services	0_Library_Universe_Service_Development	[‘Library,’ ‘Universe,’ ‘Service,’ ‘Development,’ ‘Knowledge,’ ‘New,’ ‘User,’ ‘Technology,’ ‘Intelligence,’ ‘Social Education’]
	5_Service_Public_Library_User	[‘Service,’ ‘Public,’ ‘Library,’ ‘User,’ ‘Intelligence,’ ‘Information,’ ‘Provide,’ ‘Transformation,’ ‘Universe,’ ‘Education’]
Spatial Scenarios	1_Virtual_Space_Library_Reality	[‘Virtual,’ ‘Space,’ ‘Library,’ ‘Reality,’ ‘User,’ ‘Universe,’ ‘Scenarios,’ ‘Experience,’ ‘Reading,’ ‘Entity’]
Theoretical Research	2_Universe_Library_Research_Theory	[‘Universe,’ ‘Library,’ ‘Research,’ ‘Theory,’ ‘Concept,’ ‘Discussion,’ ‘Future,’ ‘Development,’ ‘Scholar,’ ‘On’]
Smart Technology	3_Intelligence_Technology_Knowledge_Artificial Intelligence	[‘Intelligence,’ ‘Technology,’ ‘Knowledge,’ ‘Artificial Intelligence,’ ‘User,’ ‘Library,’ ‘Universe,’ ‘Intelligent,’ ‘Service,’ ‘Digital’]
Digital Resources	4_Digital_Resource_Collection_Assets	[‘Digital,’ ‘Resource,’ ‘Collection,’ ‘Asset,’ ‘Content,’ ‘Library,’ ‘Copyright,’ ‘Knowledge,’ ‘Development,’ ‘Universe’]
Digital Personas	7_Virtual_Personas_Digital_Virtual Reality	[‘Virtual,’ ‘Personas,’ ‘Digital,’ ‘Virtual Reality,’ ‘Reading,’ ‘Memory,’ ‘Real Person,’ ‘User,’ ‘Avatar,’ ‘Feeling’]
Architectural Design	8_Layer_Physical Layer_Model_Application Layer	[‘Layer,’ ‘Physical Layer,’ ‘Model,’ ‘Application Layer,’ ‘Interaction,’ ‘Software,’ ‘Architecture,’ ‘Divided Into,’ ‘Data,’ ‘Universe’]
Social Scenarios	9_Social_Alliance_Scene_Universe	[‘Social,’ ‘Alliance,’ ‘Scene,’ ‘Universe,’ ‘Resource Pool,’ ‘Learning,’ ‘Participant,’ ‘Library,’ ‘Conference,’ ‘Creation’]
Big Data Computing	10_Data_Big_Related_All	[‘Data,’ ‘Big,’ ‘Related,’ ‘All,’ ‘Integration,’ ‘Computing,’ ‘Heterogeneous,’ ‘Cloud,’ ‘Storage,’ ‘Technology’]
Maker Space	12_Maker_College_Space_Virtual	[‘Maker,’ ‘College,’ ‘Space,’ ‘Virtual,’ ‘Construction,’ ‘Drive,’ ‘Sharing,’ ‘Space System,’ ‘Innovation,’ ‘Below’]

In the positive dimension topics the “Smart Services” topic encapsulates scholars’ expectations and confidence in the metaverse enabling smart services in libraries (Chen et al. 2023; Fan 2022; Zhang and Ye 2023; Zhao and Lin 2022; Wu et al. 2022a, 2022b). Under the “Spatial Scenarios” topic numerous scholars have explored the digital-real fusion space of metaverse libraries and their spatial development trajectory, conducting surveys and designs for library spaces in various service scenarios (Wu et al. 2022a, 2022b; Guo et al. 2023; Liang and Xu 2024). In the “Theoretical Research” topic the emergence of new technological scenarios activates new academic growth points in the library field, such as the “book-person-use” axis for metaverse libraries (Zhao 2022), the “Third Space” concept of metaverse libraries (Liang and Xu 2024), and the principles of virtual-real interaction, technological interconnection, and humanistic mutual benefit (Zhao et al. 2022). Under the “Intelligent Technology” topic, the rapid development of artificial intelligence in the metaverse technology system aligns the ideal forms of the metaverse, with scholars deeply discussing its application scenarios in metaverse libraries, such as management systems (Zhang et al. 2024), intelligent consulting (Liu and Shao 2024), and intelligent retrieval (Li 2023), which are likely to be realized first. Under the “Digital Resources” topic, scholars explore services for disciplinary literature resources (Bai et al. 2023), resource construction (Hu and Ji 2023), and resource management (Li and Zhao 2022) in metaverse libraries, detailing the design, construction, and governance of information resources in metaverse libraries. Under the “Digital Personas” topic, scholars discuss the functions of users’ digital avatars in metaverse library settings (Zhang and Su 2021), the concepts of digital person services (Si and Ma 2023), and the intelligent ecosystem formed by digital persons based on artificial intelligence (Qian et al. 2023), depicting interaction scenarios shaped by user avatars and AI-based digital persons in metaverse libraries. Under the “Architecture Design” topic, scholars classify and design the constituent elements (Li and Ma 2022) and technical architecture (Li 2022) of metaverse libraries, providing reference and guidance for the application of the metaverse in smart libraries. Under the “Social Scenarios” topic, scholars contemplate new scenarios that may emerge in metaverse libraries from the perspective of the social attributes of reading promotion (Sheng 2023) and propose a knowledge exchange-based social system (Qian et al. 2023), expanding the functional attributes of library services. Under the “Big Data Computing” topic, scholars view big data technology as the technical foundation for the digital-real fusion of libraries (Zhang et al. 2022), promoting interdisciplinary integration in disciplinary support services (Bai et al. 2023; Fu et al. 2023). Under the “Maker Space” topic, scholars design a virtual maker space system architecture driven by the metaverse for university libraries (Li and Yang 2023) and propose a universal education service based on metaverse library makerspaces (Guo et al. 2022), ushering in new development opportunities for makerspaces under the empowerment of metaverse technology systems.

Additionally, the topic distance display of the positive sentiment topics before manual processing is shown in Figure 3. The size of each circle corresponds to the extraction sequence of the topics, while the distance between circles represents the degree of similarity between topics. As depicted in Figure 3, the positive dimension topics are generally clustered into three groups, primarily centered around “topic_0 Smart Services” and “topic_1 Spatial Scenarios,” “topic_3 Intelligent Technology” and “topic_4 Digital Resources,” and “topic_2 Theoretical Research.” These topics mainly focus on various service scenarios, technological architecture, and theoretical exploration, providing an essential theoretical foundation for understanding metaverse libraries in the library community.

Figure 3:

Plot of distances between positive sentiment topics.

4.2.2 Negative Dimension Topic

The results of the clustering of the negative dimension topics after manual merging and the subject entries under each subject are shown in Table 2.

Table 2:

Negative dimension topics after manual treatment.

Topic	Top 4 keywords	Keywords
Technological Development	0_Library_Universe_Research_Technology	[‘Library,’ ‘Universe,’ ‘Research,’ ‘Technology,’ ‘Development,’ ‘New,’ ‘Construction,’ ‘Intelligence,’ ‘Theory,’ ‘Service’]
User Service	1_Library_Universe_Virtual_User	[‘Library,’ ‘Universe,’ ‘Virtual,’ ‘User,’ ‘Service,’ ‘User,’ ‘Data,’ ‘Technology,’ ‘Digital,’ ‘Identity’]
Ethical Risks	2_Social Education_Empowerment_Universe_Ethics	[‘Social Education,’ ‘Empowerment,’ ‘Universe,’ ‘Ethics,’ ‘Risk,’ ‘Library,’ ‘Scene,’ ‘Fairness,’ ‘Equality,’ ‘Coping’]
From Reality to Virtual	3_Universe_Library_Not_Lack	[‘Universe,’ ‘Library,’ ‘Not,’ ‘Lack,’ ‘Complete,’ ‘From Reality,’ ‘Platform,’ ‘Project,’ ‘Shanghai Library,’ ‘Virtual’]
User Interface	5_User Interface_Design_User_Retrieval	[‘User Interface,’ ‘Design,’ ‘User,’ ‘Retrieval,’ ‘Graphic Interface,’ ‘Communication,’ ‘Area,’ ‘Browse,’ ‘Access,’ ‘Should’]
Resource Development	6_Resource_Data Model_Development_Again	[‘Resource,’ ‘Data Model,’ ‘Development,’ ‘Again,’ ‘United States,’ ‘Proper,’ ‘Element,’ ‘Professional,’ ‘Data,’ ‘Library’]
User Privacy	8_User_Privacy_Information_Will	[‘User,’ ‘Privacy,’ ‘Information,’ ‘Will,’ ‘When,’ ‘Data,’ ‘Trigger,’ ‘Only,’ ‘Resource,’ ‘Sensory’]

Under the “Technological Development” topic many scholars express concerns about constructing metaverse libraries due to the lag in technological development. Some argue that libraries should be cautious about potential social issues and uncertainties while keeping pace with the development of metaverse technology (Dong and Bu 2022). Additionally, challenges arise in constructing metaverse libraries related to the increased technical skill requirements for librarians and the potential inequalities that users face (Li et al. 2023). In the “User Services” topic it is suggested that merely replicating the service system of physical libraries in virtual environments may lead to low reflexivity (Niu et al. 2023). The development of the metaverse not only involves the digitization of libraries but also necessitates a transformation of service systems, requiring a redesign of the service framework (Li and Ma 2022). Furthermore, measures should be taken to ensure the accuracy and authenticity of users’ digital identities during user services (Lu et al. 2023a). Regarding the “Ethical Risks” topic, some opinions suggest that in metaverse libraries, where users appear as virtual avatars, the boundaries of gender, ethnicity, community, and other constraints are broken, leading to certain ethical risks (Tang 2023). Digitalizing users in metaverse libraries blurs the boundaries of function, identity, and value, thereby generating ethical risks (Huang 2023). Under the topic of “From Reality to Virtuality” several scholars express concerns about the risks of moving from reality to virtuality during the development of the metaverse (Chen et al. 2022; Zhao and Lu 2022; Cai et al. 2023) as this may lead to the neglect of the construction of physical collections and services in libraries to some extent. Regarding the “User Interface” topic, some viewpoints suggest that the interaction interface design of auxiliary facilities in library metaverses should be more concise to accommodate the usage experience of every user (Zhou et al. 2023). In the “Resource Construction” topic some opinions highlight that during the construction of resources in metaverse libraries, attention should be paid to issues such as intellectual property rights and user privacy (Hu and Ji 2023). Additionally, it is noted that during interlibrary collaborative resource construction attention should be paid to technological gaps between libraries (Yang et al. 2021). Under the “User Privacy” topic, some viewpoints assert that while the metaverse empowers the intelligent transformation of libraries it also brings issues of data privacy infringements for “present” participants (Niu et al. 2023). Librarians can explore open-source software like Reaction Grid to mitigate privacy and security risks users face in the metaverse (Xing et al. 2023).

Through the display of topic distances for the negative sentiment topics before manual processing, as shown in Figure 4, it is evident that negative topics in metaverse libraries generally cluster into two groups, primarily consisting of first “topic_0: Technological Development” and “topic_2: Ethical Risks” and second “topic_1: User Service.” Negative topics primarily center around the risks brought about by integrating technology in the metaverse and libraries and the challenges libraries face in innovating their service systems. In discussing negative dimensions, the academic community conducts forward-looking explorations into potential issues in the development, construction, and management processes of metaverse libraries, aiming to provide warnings and insights for future metaverse development.

Figure 4:

Plot of distances between negative sentiment topics.

5.1 Synthesis of Sentiment Topics

Metaverse libraries will likely be at the next stage of developing smart libraries. Through sentiment-topic mining of related research on Chinese metaverse libraries it is found that current research holds positive attitudes towards aspects such as smart services, spatial scenarios, theoretical research, intelligent technologies, digital resources, digital personas, architectural design, social scenarios, big data computing, and makerspaces. However, there are concerns regarding technological development, user services, ethical risks, departure from reality, user interfaces, resource construction, and user privacy. It is noteworthy that scholars present different viewpoints regarding service construction-related topics. This does not seem contradictory; rather, positive and negative sentiment topics complement each other. In the positive sentiment topics of smart services, AI-based navigation, retrieval, consultation, recommendation, and other services are widely supported by scholars as smart service application scenarios (Chen et al. 2023; Li 2022). Interactivity, immersion, and intelligence are the service characteristics of metaverse libraries (Tang 2023), wherein providing services with these features through constructing virtual venue spaces has become a common approach (Kong 2020; Yang et al. 2023). This aspect is supplemented by other scholars in negative sentiment topics, stating that the services provided by metaverse libraries should not simply replicate the service system of real-world libraries (Dong and Wu 2023) but require a redesign of the service system (Li and Ma 2022). This also demonstrates that sentiment-topic sequential literature analysis can more clearly analyze different viewpoints under a single topic, thereby better understanding the development of metaverse libraries.

The spatial construction of metaverse libraries serves as the environmental carrier for intelligent services. Discussions on spatial construction primarily focus on the future metaverse libraries scenarios design, such as the construction of makerspaces based on virtual reality technology (Li and Yang 2023), interdisciplinary virtual knowledge exchange spaces (Fu and Shao 2024), and learning spaces that utilize virtual reality, artificial intelligence, and other technologies to enhance user learning efficiency (Bai et al. 2023). Integrating virtual and real elements is a fundamental characteristic of metaverse libraries (Wu et al. 2022a). In the context of the library this integration not only involves using technological means to merge virtual and real environments but also entails collecting real-world data and employing simulation and big data-based predictive analytics to understand users better, thereby providing user-centered services.

Supporting the content and services scenarios are metaverse-related technologies, “VR space integration” and “fundamental technologies” serving as the main driving factors, while “object connectivity” and “communication computing infrastructure” play a supportive role (Tukur et al. 2024). However, in the process of technology application, issues such as the ethical considerations of virtual communities, protection of virtual assets, dissemination of false information, and user privacy breaches, which have existed in the development of the Internet, are also worthy of attention in the construction of metaverse libraries (Di Pietro and Cresci 2021; Falchuk et al. 2018; Wang et al. 2023). Therefore, exploring governance schemes for metaverse libraries may help prevent these issues. External legal constraints and internal technological security controls are essential for governing metaverse libraries (Lu et al. 2023b; Wen 2023). Standardized digital content accountability, along with precise and rapid decision-making in response to emergencies, plays a significant role in the governance process of metaverse libraries and is crucial in the governance of the digital society (Zhao and Lu 2022).

In the literature on metaverse libraries, various research topics rarely appear in isolation. As demonstrated in the research results above, the positive-dimensional topics mainly cluster into three groups, while the negative-dimensional topics mainly cluster into two groups. Taking the positive-dimensional topics as an example, when discussing the intelligent services of metaverse libraries, the library’s space and scenarios often appear in the same discourse field of inquiry as the medium for providing services. Meanwhile, intelligent technology and digital resources are prerequisites for providing smart services in metaverse libraries. The former serves as the supporting technology for service provision while the latter constitutes the service content itself. Both appear together in texts discussing metaverse libraries’ construction strategies and measures.

As stated in this paper’s research significance, the research methods employed can provide a better understanding of the concept of metaverse libraries and related studies. On the one hand, it enables people to understand what scholars are discussing regarding metaverse libraries and which aspects are viewed with hope or concern. On the other hand, it allows for integrating different sentiments and perspectives under the same topic, facilitating a better understanding of the construction and development of metaverse libraries. Moreover, during the thematic mining process, we can better understand the logical connections between topics and gain a clearer understanding of content in metaverse library literature.

5.2 Prospects for Future Research

With the rapid advancement of large language models, artificial intelligence (AI) has emerged as the leading technology within the core technological framework of the metaverse, particularly in supporting the vision of building a metaverse library. Leveraging large language models, libraries can design more convenient and intelligent technology use scenarios, such as AI-assisted teaching support services (Okunlaya et al. 2022), book cataloging (Abutayeh 2024), and librarian training (Lo 2024). Currently, the application of AI services in libraries remains limited, possibly because libraries are still in the evaluation stage of innovation and still need to be in the confirmation phase (Harisanty et al. 2023). There are concerns that large-scale applications of AI might replace existing librarian roles. Studies have addressed this concern and investigated the perspectives and opinions of library leaders and librarians regarding AI (Harisanty et al. 2024; Hervieux and Wheatley 2021; Cox et al. 2019). However, it is crucial to consider whether AI can genuinely meet the urgent needs of library patrons and provide more satisfactory services than human librarians. Further research could consider starting from the needs of readers and their service experience in libraries, using case studies and self-reported data from readers to assess the necessity of large-scale AI adoption in libraries. This assessment could help advance the application of AI technology in libraries, ushering in the era of smart libraries.

Preventive Measures for Potential Issues. According to the research results above, the academic community has raised concerns regarding the application of metaverse-related technologies in libraries, including ethical issues such as privacy breaches and misinformation. Personal information is fundamental to enhancing user satisfaction in AI-based services; without the use of personal data, the effectiveness of project recommendations would significantly diminish (Sun et al. 2024). Therefore, the primary issue regarding privacy is not whether to disclose information to library systems; rather, the scope and ownership of information usage need to be clearly defined. If information is leaked beyond the user-permitted scope there should be methods to compensate the user. This process can be viewed as the personal asset management of user data. Research has designed a library smart value-added service platform based on blockchain technology (Hu 2022) to ensure the security of information collection, storage, and dissemination. However, further studies are needed to explore the ownership and value of user data in future smart libraries. Regarding the issue of misinformation, it is crucial to determine the technological path for constructing smart libraries based on metaverse-related technologies. The two main paths are virtual reality, such as libraries in Second Life, or a mixed reality approach, where intelligent services are experienced in real library settings using AR headsets. In the mixed reality approach, readers do not have virtual identities, which can prevent unethical behavior from separating virtual and real identities. Further research could compare and evaluate the feasibility of these two technological paths from multiple perspectives, including technology implementation, user needs, service effectiveness, and design corresponding construction methods.