A Bibliometric Analysis of Virtual Influencers as a Trend in Digital Marketing: Insights from WOS

The concept of virtual influencers is becoming increasingly important in today's world. Especially when it is considered that by January 2024, more than 5.04 billion users of social networks existed which represented 62.3% of the world's population (DataReportal, 2024) and the numbers are increasing daily. With the increase of social media usage, the number of virtual influencers has also expanded significantly in recent years (Grand View Research, 2024). However, the question arises as to how and why the concept of the virtual influencer has emerged (Um, 2023a). The key to the success of virtual influencers lies in live streaming, but also several limiting factors related to the fact that influencers are human beings who need to fulfill basic life needs such as the need to eat, drink, sleep, socialize in the physical environment and other accompanying activities that are not closely related to their work should be pointed out. It is precisely from these limiting factors that the need for companies to create virtual influencers arose, which very quickly became an integral part of the digital marketing mix in the form of branding (Appel et al., 2020).

Social networks have become an integral part of daily lives, satisfying social, information, and entertainment needs. So, platforms like Instagram, TikTok, Twitter, and Facebook allow users to publish their content, stay informed about others, and follow trends, brands, and celebrities - from athletes and actors to singers. According to that, a new category of celebrities has emerged in this digital landscape, and they are called influencers. These individuals are experts who carefully curate their public image through personal branding, which becomes visible through their online presence (Arsenyan and Mirowska, 2021a). In general, influencers are particularly used in the context of social media marketing. They can be hired by marketers and companies to represent and promote brands, but they can also act independently to achieve specific personal goals (Belanche et al., 2024a). However, several researchers confirm that the credibility and authenticity of the content they share are being questioned (Hrustek, 2023). Many companies saw their opportunity in virtual influencers who are not real people, but digitally created influencers who maintain their influential status through consistent imagery and persuasive effectiveness based on the functionality of humans in combination with audio-visual characteristics (Arsenyan and Mirowska, 2021a).

Virtual influencers are computer-generated individuals designed to attract as many followers as possible on different online platforms (Koles et al., 2024a). In other words, they are avatars based on artificial intelligence and designed with the help of computer graphics software. Behind their work is a team of experts and digital agencies who have ensured that the virtual influencers take on a soul, i.e. human traits and develop a lifestyle characteristic of real people. All the above steps are carefully designed to capture the interest of the potential audience and keep them on social networks for as long as possible, so that potential consumers have time to become aware of the importance of the brand offered by the company (Koles et al., 2024a). The first computer-generated social media virtual influencer on Instagram is Miquela Sousa, also known as Lil Miquela, created in 2017 by Brud media agency. Miquela has the characteristics of a “real human” but she is neither a human person nor a robot (Robinson, 2020a). It is important to highlight that the concept of virtual influencers is gaining increasing prominence as a trend in digital marketing, driven both by advancements in digital technology and by their ability to offer followers benefits comparable to those of human influencers. (Lim and Lee, 2023a). But also this concept is still a relatively new and unexplored, but rapidly growing and very important field (Um, 2023a) which should be further explored especially because the virtual version of influencers can threaten the effectiveness of human ones and replace them.

The upcoming chapter introduces the theoretical background for conducting bibliometric analysis on virtual influencers as a prominent trend in digital marketing. The third chapter presents the research methodology, followed by the fourth chapter, which shows the research findings, including descriptive statistics, visual analyses, research implications, and limitations. The paper concludes with a summary of the conducted research.

In recent times, people's lives have been engulfed by intriguing changes in human behavior, especially among generations Z and Alpha. These tech-savvy generations have seamlessly integrated social networks into their lives from an early age. However, a notable challenge arises from the blurring of boundaries between the virtual and the real world (Choi et al., 2023a). Many individuals within these generations do not recognize the differences between online experiences and offline reality (Chiu & Ho, 2023a). In response to this evolving landscape, companies are adapting their marketing strategies to this generation of consumers. One of the most intriguing trends is the development of virtual influencers. As mentioned in a paper by Robinson, this term refers to computer-generated personalities who wield influence across digital platforms.

Virtual influencers represent computer-created and designed digital entities that look completely like people, which are integrated into human networks and which communicate with their users (Mirowska & Arsenyan, 2023a), (Stein et al., 2022a) to achieve certain higher goals of those who created them (Koles et al., 2024a). These virtual influencers assume various roles and perform distinct tasks when interacting with people, adapting to the context and goals at hand.

With the expansion of digital spaces, the role of virtual influencers is becoming increasingly important. Although virtual influencers are fictional characters, they arouse interest and attention among users due to their attractive appearance, human characteristics and the way they communicate with them on social networks (Na et al., 2023a). The research paths pursued by authors in this field can be categorized into several fundamental groups. Most authors deal with the influence that virtual influencers have on their followers. In this context, the authors examine the feelings, relationships and interactions of consumers (Stein et al., 2022a), (Mirowska & Arsenyan, 2023a), (Jhawar et al., 2023a), (Liu & Lee, 2022a), (Na et al., 2023a). Some authors devote themselves to researching parasocial relationships with virtual influencers by investigating whether consumers develop feelings, empathy or emotional attachment to virtual influencers (Melnychuk et al., 2024a). In their research, certain authors have highlighted ethical challenges and expressed apprehensions regarding virtual influencers. They advocate for a heightened awareness of moral responsibilities throughout this process (Robinson, 2020a), (Conti et al., 2022a), (D. Kim & Wang, 2023a). As the role of virtual influencers is often seen in the context of content marketing (Suprawan & Pojanavatee, 2022a), (Q. Zhou et al., 2024a), some authors investigated the influence of virtual influencers on purchase intent and purchase decisions (H. Kim & Park, 2024a) i.e. consumer behavior based on purchasing decision-making process. Furthermore, research papers show that the recommendations of virtual influencers are positively related to purchase intent (El Hedhli et al., 2023a). Research also indicates that virtual influencers can engage consumers as effectively as human influencers, however, their artificial nature raises concerns about credibility, potentially undermining the effectiveness of marketing efforts (Lim & Lee, 2023a), (Chiu & Ho, 2023a).

The penetration of virtual influencers into human networks is particularly significant on social media platforms (Stein et al., 2022a) where they often act as an alternative person in brand advertising (Liu & Lee, 2022a) and virtual ambassadors in projects and promotional activities (Zelenskaya & Rundle-Thiele, 2022a), (Xie-Carson, Benckendorff, et al., 2023a). Virtual influencers are often used for promotion in tourism (Xie-Carson, Magor, et al., 2023a), (Zelenskaya & Rundle-Thiele, 2022a), but also in some other domains. For example, Donald Trump as one of the candidates for the president of the United States of America in 2024, used influencers and images created by artificial intelligence in his campaign (Sainato, 2024).

Despite the emergence of virtual influencers in digital marketing, several unanswered questions persist. First, researchers are still exploring how audiences in digital environment interact with these artificial digital personalities. Additionally, some experts caution against potential risks associated with using virtual influencers in marketing communication connected to ethical issues. So, it should be emphasized that the need for further investigation of this concept certainly exists.

The main goal of this paper is to give a detailed review of the literature published until April 18th, 2024, which dealt with the research of virtual influencers as a current trend in digital marketing. To fulfill the research goal, desk research was conducted.

This paper presents a bibliometric analysis based on the collected data, applying a quantitative and systematic method for identifying, evaluating, and monitoring published papers (Sindhu & Bharti, 2020), which consists of statistical and mathematical methods applied to sets of bibliometric references in scientific literature on a certain topic (Lacković Vincek et al., 2023). Standard bibliometric analysis, combined with descriptive statistics, serves to identify authors, their connections, keywords, citations, journals, organizations, countries, and other criteria, with the aim of uncovering their relationships in a scientific context (Lacković Vincek et al., 2023).

As is already known, data for bibliometrics can be obtained using various databases, including Scopus database and Web of Science (WoS) platform. According to the previous research, bibliometric analyses have been conducted on topics related to this research, such as general concepts of influencers in connection with social networks or digital marketing, but not in the context of virtual influencers. However, as far as the specific bibliometric analysis of virtual influencers is concerned, only two papers were published in the aforementioned relevant databases. One paper, visible in the Scopus database, was published by the author Barbosa in 2024 (Barbosa, 2024). Another, visible in the Web of Science core collection database, was published by the authors Vila López, Kuster Boluda, Mora Pérez and Pascual Riquelme (Vila-Lopez et al., 2023a). However, in their case, the same search query was not used to search the Web of Science core collection database, and also used visualization tool was not the same. Unrelated to the title of the paper itself, the aforementioned paper focuses on the research of the influencer concept, although in some places the concept of virtual influencers is referred to as well. Therefore, based on insight into the current state, the authors decided to create a bibliometric analysis that includes only the Web of Science core collection database. Considering that this kind of paper in the mentioned database does not exist so far, the contribution of this paper is evident.

So, for the purposes of writing this paper, the WoS database was considered suitable for several reasons. First, the goal was to reduce the possibility of repeating the analysis of the same studies because the same paper can be in several databases, especially in the Scopus and WoS databases, and this can go unnoticed (Mavric et al., 2021). In addition, WoS database was chosen because it is considered the most selective database (Singh et al., 2021) with the highest impact factors and citation analysis, and is one of the world's leading relevant databases (Pavlić, 2024).

As mentioned earlier, the literature on virtual influencers published until April 18, 2024, was scanned in the Web of Science core collection database. The search term applied to identify the closest matching publication included the search query “virtual influencer” OR “AI influencer” OR “artificial influencer” AND “digital marketing”. The search covered all variations of the relevant terms across all fields. For each document that met the criteria, information such as publication year, language, journal, title, author, affiliation, keywords, document type, and citation count was collected and exported in plain text format. The VOSviewer software (version 1.6.10) was then used to analyze co-authorship, co-occurrence, citations, bibliographic coupling, co-citations, and research topics.

To summarize, a bibliometric analysis was used to show the evolution of virtual influencers as a trend in digital marketing. The research process consisted of the following stages:

Research design – As authors through the years were researching the concept of social networks and digital marketing in general, a new trend in scientific papers since 2020 appeared. That term was a virtual influencer and it seemed that it needed more research. So, a bibliometric analysis is the first step for the authors to go deeper into this new research field.

The concept of virtual influencers as a recognized trend with the development of technology includes the integration of the traditional concept of influencer marketing and social networks as one of the elements of the promotional mix of digital marketing. This is also evident from the analysis of the most frequently used keywords in published papers. Based on the examination, the keyword analysis and the main research areas related to virtual influencers, it is evident that there is an impact between the mentioned topics. Since this field combines business and economics with information and communication sciences, as well as psychology, it is evident that virtual influencers represent an interdisciplinary trend that is becoming increasingly important. Although the number of papers is still small, and only a limited number of authors, institutions, and countries are involved in this research area, it is expected that this will change in the future. Until then, the scientific contribution of this paper remains significant.

However, this research has certain limitations. The main limitation is that the bibliometric analysis relies solely on the Web of Science core collection database. Although Web of Science is comprehensive, future studies should include additional scientific databases to provide a more complete analysis and enable comparison of results across different sources The limitation of the research can also be found in the set search query, because in this way the authors limited themselves.

Digital technologies, including artificial intelligence, are transforming marketing practices. The boundary between real and digital life is blurred, and people now utilize every available moment to engage in the digital world. As a result, marketing agencies and brand companies are adopting agile approaches to create new content for the social digital environment. In recent years, social networks have become a significant digital space, dominated by influencers. These individuals intertwine everyday life, activities, and brand-related content to shape the awareness and habits of their audience. Innovative directions resulted in new trends in digital marketing and one of example is computer-generated virtual influencers who interact with their users, sharing experiences as human influencers. Interestingly, scientific society began recognizing virtual influencers around 2020 and the first one was created three years before, in 2017. Notably, two authors stand out and they are Robinson, who was the first to delve into this phenomenon, and Arsenyan, the most frequently cited author in this field. It's worth noting that this intriguing topic has captured the attention of several other scientists. The results of the bibliometric analysis point to the fact that virtual influencers are an under-researched term in digital marketing. However, this will certainly change in the future, as it is believed that the use of virtual influencers will increase in other areas of information and communication sciences as well. The visualization findings indicate that the concept of virtual influencers is closely linked to social media and influencer marketing. Additionally, based on the keywords, it becomes evident that this topic revolves around the impact that virtual influencers exert on their followers or consumers, influencing their behavior. Moreover, researchers are presently re-evaluating the dynamics between virtual influencers and users, highlighting the dubious authenticity, deficient empathy, and questionable credibility of virtual influencers. These issues are expected to become more pronounced, considering the emerging trend of deinfluencing. According to the findings from this paper, future research directions will be focused on expanding this research with additional literature from other relevant sources. In future research, the overview of the field will be expanded through the identification of key research areas and suitable techniques/methods in the implementation of research. Some of the “hot topics” that are in this paper recognized as important in the field of virtual influencers and will occupy future attention can be divided into three key directions: (1) the psychological effects of interacting with virtual influencers, (2) the authenticity and credibility of virtual influencers, (3) technological aspects, regulation and ethical issues for virtual influencers. These main three directions can serve as a foundation for future research, while always considering the ongoing advancements in the technology itself.