Abstract
This case study examines how and why people follow Rozy, a prominent virtual influencer in South Korea, and how her followers perceive her in comparison with human influencers and celebrities. Drawing on online observation of Rozy’s Instagram account and in-depth interviews with 10 followers, the study explores the relationships that emerge around a virtual influencer in practice. The findings show that following Rozy did not necessarily indicate fandom. Rather, the participants in this study often followed her out of curiosity about digital trends, technological novelty, and visual style, while maintaining relatively low levels of affective attachment. Some visible follower accounts also consisted of virtual, corporate, or unidentifiable accounts, complicating any simple equation between follower counts and fandom. Although participants sometimes viewed Rozy’s virtuality positively—for example, as freeing her from the scandals associated with human celebrities—her limited real-time interaction, lack of embodied authenticity, and thin narrative depth constrained emotional attachment and trust. Participants also contrasted Rozy with animated or fictional characters who invite deeper affection through richer backstories and greater interpretive openness. Based on these findings, the study conceptualizes Rozy’s followers, in this case, as closer to “observers” than to fans or consumers. The article argues that visual realism alone is insufficient to sustain fandom around virtual influencers; rather, interactivity, credibility, and narrative richness are more important in shaping audience attachment.
Keywords
Introduction
In 2021, a woman who appeared to be in her twenties featured in a television commercial for a major South Korean bank. It was unusual for her to be a newcomer in Korean media, as bank advertisements typically showcase mature, reliable celebrities familiar to the audience. In contrast, this new figure promoted an insurance product by dancing to upbeat music. Following the commercial, her management agency secured an exclusive contract with the bank, and she was later named one of South Korea’s representatives for attracting the 2030 Busan EXPO. While such celebrity endorsements may seem routine to the public, Rozy stands out: she is not a human but a virtual influencer.
Although most Korean media coverage refers to Rozy as a virtual “celebrity” (e.g., Oh and Choi, 2021), her appearance in this commercial represents a significant shift in casting practices from the perspective of celebrity studies. The presence of fans has long been a key indicator of celebrity status, as it is celebrities’ ability to connect advertisers with an existing fan base that justifies their hiring (Dyer, 1979, 1986; Shahin and Hou, 2025; Turner, 2014). Today, this concept extends beyond traditional celebrities in mass media; influencers, who gain visibility through social media and foster smaller yet highly dedicated fan communities, occupy a similar role in the modern celebrity landscape.
In Rozy’s case, however, her Instagram following was around seventeen thousand—barely a number that would typically classify her as a celebrity. She also did not fit the mold of a typical influencer, who usually engages in ongoing interaction with followers and has established a successful business based on a strong fan base, as many South Korean influencers have done. Despite lacking significant achievements in conventional entertainment or social media, Rozy has garnered considerable media attention and is often referred to as a virtual “celebrity” (e.g., Oh and Choi, 2021). Her Instagram bio even labels her as Korea’s first virtual “influencer.” Following a bank commercial, Rozy’s follower count surged to 17 thousand by June 2025.
In these respects, the present study situates virtual influencers within the broader scholarship on celebrities, including SNS influencers, while examining how they differ from existing forms of celebrity. Focusing on the case of Rozy, this study addresses the following questions: (1) In the case of Rozy, how and why do people choose to follow a virtual influencer? and (2) What attributes do followers perceive Rozy to possess in comparison to human influencers or celebrities?
Theoretical backgrounds
From celebrity to influencer: Platformized fame and relational consumption
In modern societies, including South Korea, buying products through social media influencers has become a common way of consuming. Instead of discovering products primarily through television commercials or print ads, users now find and assess goods through the daily posts, vlogs, and recommendations of social media figures known as “influencers” (Feng et al., 2021; Lou et al., 2019). Influencers are individuals who, while not necessarily widely recognized as “celebrities” in the traditional sense (Turner, 2014), maintain authentic connections with a niche group of followers on social media. They hold a visible presence within specific interest communities or taste cultures on digital platforms.
Historically, fame and the dissemination of images centered around celebrities. According to Turner (2014), since the early twentieth century, when American cinema began to thrive, celebrities have been crucial to the evolution and commercialization of popular culture. In this system, celebrities are recognized for their achievements in fields such as acting, music, or sports, and their fame is rooted in these accomplishments (Dumitriu, 2018; Gamson, 2023). The public, in turn, invests desire and fantasy in the media portrayals of these individuals, linking celebrity fame closely to the public’s interest in media productions featuring them.
Traditionally, mass media (e.g., television, magazines, film, or radio) served as the primary avenues for the public to access celebrities, shaping the construction and circulation of fame. However, with the emergence of social media, ordinary users of digital platforms can now achieve visibility and fame comparable to that of traditional celebrities. These individuals are often referred to as “micro-celebrities” (Abidin and Ots, 2016; Khamis et al., 2017; Senft, 2008) or “DIY celebrities” (Turner, 2014), highlighting that they are not necessarily institutionally managed. Instead, they curate their self-presentation, manage their own audiences, and engage directly with followers. Unlike traditional celebrities, whose fame is built on accomplishments within specific fields, influencers gain recognition by consistently producing content that aligns with a self-defined brand (Senft, 2013). Accordingly, Piehler et al. (2022) distinguish between “original social media influencers,” including “instafamous” figures who gained fame exclusively through social media, and traditional celebrities who initially garnered recognition outside of social media and later expanded their presence to these platforms (Brooks et al., 2021; Hess et al., 2022).
The maintenance of an influencer’s status critically relies on how they stage and sustain relationships with their audiences. Influencers’ interactions extend beyond personal connections, such as friends or family, to a wider audience of followers who like, share, or comment on their content (Piehler et al., 2022). To gain and maintain visibility and attention, influencers continuously post content that reflects their everyday lives, ranging from the trivial and mundane to exciting snippets of exclusive opportunities in their work (Abidin and Ots, 2016: 154–155). This ongoing stream of posts and stories invites followers to observe what seems to be the influencer’s daily routine, minor challenges, consumption choices, and special events.
Historically, the relationship between celebrities and fans has been driven by desire for a distant, idealized persona. In contrast, the dynamic between influencers and their followers is distinct. Influencers create a sense of closeness, relatability, and accessibility (Xu et al., 2025). By sharing curated glimpses of their everyday lives and engaging in direct communication, influencers foster an emotional and seemingly reciprocal relationship with their followers (Alperstein, 2019). This approach allows influencers to build personal brands rooted in intimate connections, which companies and advertisers can leverage for consumer outreach (Hearn and Schoenhoff, 2016: 194).
Virtual humans often exhibit characteristics similar to traditional celebrities; their appearance, backstory, and brand partnerships may resemble those of entertainers or fashion models. However, they also adopt the posting styles, interactive practices, and intimate tones typical of influencers (Byun and Ahn, 2023; Chiu and Ho, 2023). Thus, virtual influencers exist in a hybrid space, merging the aspirational allure and spectacle of celebrity culture with the perceived closeness, ordinariness, and continuous interaction associated with influencer culture. This dual nature prompts questions about how followers perceive their relationship with these entities, how they assign agency, emotion, and responsibility to virtual humans, and how they define their roles as fans, consumers, or merely curious observers of technological innovation.
Studies about virtual humans
Studies on virtual humans, including virtual influencers, can be categorized into two main strands: (1) The first strand focuses on the inherent artificiality of virtual humans and emphasizes technological efforts to realistically replicate human appearance and behavior; (2) The second strand, often grounded in cultural studies, examines the subjectivity and agency of virtual humans.
The first strand primarily explores virtual humans in the entertainment industry, a space traditionally occupied by human celebrities. These studies highlight the technological aspects that contribute to the perceived realism and naturalness of virtual humans, reflecting advancements in technology. For instance, Kim et al. (2023) trace the evolution of Korean virtual idols over three decades, particularly addressing how the “uncanny valley” phenomenon has been navigated. Similarly, Hwang and Lee (2021) investigate audience emotional responses to virtual humans in advertising as influencers, while Hong and Park (2020) examine techniques for effectively simulating spatial environments to enhance the realism of virtual humans. Within this strand, studies typically conceptualize the relationship between virtual influencers and their followers as primarily marketer-consumer oriented. For example, researchers have conducted surveys on the relationship between perceived anthropomorphism and satisfaction with virtual influencers (Park and Han, 2024), compared the influences of human and virtual influencers (Lee et al., 2025), and analyzed brand attitudes and purchase intentions related to virtual influencers (Lee and Lee, 2023).
The second strand of research is based on cultural studies perspectives emerging since the 2020s, emphasizing the subjectivities and identities of virtual humans. For example, Kim and Han (2020) analyzed the virtual K-pop idol group K/DA, based on characters from the game League of Legends, and concluded that a new digital subjectivity exists in a third space beyond traditional binaries such as virtual versus real, information versus material, and machine versus human. They also examined virtual YouTubers, noting how these creators increasingly emphasize their distinct identities by showcasing their non-human traits. In a similar vein, Kang and her colleagues (2023) explored the storytelling techniques of a virtual influencer on Instagram, revealing that the influencer’s virtual nature does not hinder their appeal. Furthermore, Park and Han’s (2024) analyzed the virtual idol PLAVE, illustrating how its virtuality emerges from a dynamic interplay of surreal elements, realistic information, non-factual imagery, and real human bodies.
In summary, while earlier research primarily aimed to make virtual humans visually indistinguishable from real ones, more recent studies have shifted to exploring the complex dimensions of virtual beings. However, the initial research approach is still limited by its marketing-focused perspective, which views the relationship between virtual humans and their followers mainly through the lens of consumer dynamics. On the other hand, the second line of inquiry recognizes virtual humans as autonomous agents but fails to adequately address the fundamental question of what types of relationships might exist between virtual humans and their followers. Consequently, this study will address the following research questions: (1) In the case of Rozy, how and why do people follow a virtual influencer? (2) What attributes do these followers perceive Rozy to possess compared to human influencers or celebrities?
Methods
This study first conducted online observations of the Instagram account 1 of the virtual influencer Rozy, followed by in-depth interviews with 10 of Rozy’s followers. During the initial stage of online observation and interviews, we continuously revised our questions, ultimately settling on the three research questions described in the previous section. The process unfolded as follows.
In the initial stage of the study, we posed questions based on the assumption that those who follow virtual humans are their fans: “How do fans of virtual humans become fans?” and “Why do they become fans?” To investigate these questions, we primarily employed two qualitative methods—online observation and in-depth interviews—for triangulation (Denzin, 2012). Our focus was on Rozy, the most prominent virtual human in Korea, who has 1.71 million followers. These assumptions were particularly relevant to Generation Z.
We began by observing Rozy’s follower list and her interactions on Instagram from January to March 2023. Due to Instagram’s policy restricting public access to complete lists of followers for verified accounts (indicated by blue check marks), we examined the first 48 displayed follower accounts and revisited these followers’ profiles monthly. Additionally, we explored the accounts of those who commented on Rozy’s posts. This approach allowed us to analyze profile pictures, descriptions, and content, helping us understand the followers better. Through this process, we discovered a distinction between Rozy’s followers and her actual fans, indicating that the two groups are not necessarily the same. We also identified accounts owned by virtual humans or companies for promotional purposes. This finding prompted us to refine our final research questions.
From March to April 2023, we conducted in-depth interviews with nine women and one man in their twenties who either followed Rozy or expressed interest in her Instagram account. These interviews aimed to explore why and how participants interact with (virtual) influencers on the platform. While there may be age variations among Rozy’s followers, we considered young adults in their twenties as a representative group. Rozy is designed to engage Generation Z, as highlighted in promotional articles (Dong, 2022). More broadly, young adults in their twenties, often referred to as “digital natives” (Prensky, 2001), actively engage with social media influencers and tend to trust their recommendations when making purchasing decisions (Park et al., 2021).
Participants were recruited through two online university communities in Seoul. We posted a recruitment advertisement in the online community “Every Time,”
2
which is widely used by university students in South Korea. The advertisement read: “We are a research team studying the virtual human Rozy. We are looking for participants who follow Rozy’s account on Instagram and regularly use Instagram to take part in an interview.” Although recruitment occurred within the community, not all participants were university students; there were nine students and one job seeker. Each interview lasted one to 2 hours and took place in either a classroom or a coffee shop. In the initial stage of the study, the interviews primarily focused on three groups of questions: (1) The characteristics of followers of virtual influencers
3
: How and why they use Instagram; Who they are interacting with on Instagram, including with Rozy; What main purposes or motivations for using Instagram, etc. (2) Interactions between social media users in their twenties and celebrities/influencers: How and why (or not) do they follow influencers on Instagram and interact with them? (3) Interactions between followers and virtual influencers: Why they follow Rozy; since when they’ve known about Rozy; why they (do not) follow virtual humans; whether there are differences between following and interacting with virtual humans and human influence, etc.
While the interviews were ongoing, we added the following questions based on insights gained during the discussions. For instance, since participants indicated that they purchase products from human influencers because they see real bodies using those products, we included a question about their intention to buy a product that Rozy might sell. Additionally, we asked interviewees whether they would feel more comfortable if Rozy resembled a human in terms of physical appearance, as several participants expressed that Rozy feels too artificial, leading to discomfort. Finally, some participants mentioned their attachment to non-existent cartoon characters, prompting us to inquire about the possibility of similar affection toward Rozy and the factors that could foster such feelings.
The procedures of making comparisons, posing questions, and sampling based on evolving theoretical concepts are essential features of this study. From the interviews, we categorized participants’ characteristics, motivations for following Rozy, modes of interaction, and attitudes toward Rozy by comparing and identifying properties and dimensions in their responses (Glaser, 1992). Additionally, we analyzed Rozy’s Instagram account as a method of triangulation to validate our findings.
Results
Reasons why we follow Rozy
With my second account, following Rozy is not a big deal
As explained in the previous section, Rozy’s followers are not necessarily her fans. The participants followed Rozy, but how do they form a relationship with her? Through the interviews, every participant mentioned that they have at least two Instagram accounts. Furthermore, all participants with multiple accounts followed Rozy from what they considered a less meaningful account for social and personal reasons.
For instance, Participant F, who has two Instagram accounts, stated that her primary account is carefully curated, with a selective list of who she follows and who follows her. In contrast, her second account is used to follow any account, such as a restaurant’s account to receive a complimentary dessert. In Participant F’s case, she follows Rozy from her second account. Similarly, Participant I follows Rozy from an account dedicated to attractive movie stars, girl groups, and celebrities. Another participant, H, described how she utilizes her second account as follows. I primarily use my second account to access information that aligns with my interests. For example, I'm particularly passionate about music, especially rock, so I follow various accounts that introduce relevant festivals and events. I mainly utilize it as an information source. (Approximately how many such accounts do you follow?) More than 200. (H)
Digital natives consider their online social relationships as important as those offline, and they view practices on social media—such as following someone or being unfollowed by friends—as carrying significant social weight. In the interviews, some participants exhibited these traits; for instance, Participant B stated, “I would really dwell on who unfollowed me and feel quite upset.” However, participants with second or third accounts noted that the social significance of their accounts can vary based on their intended purposes. They often maintain multiple accounts, typically distinguishing between private, exclusive accounts and more open, less significant ones. As Participant H explained, following someone on their secondary accounts holds less meaning compared to following someone through their primary accounts, which are used for interacting with close friends or socially significant individuals. This indicates that following Rozy through participants’ second or third accounts lacks deep meaning or significance.
We want to know what is trending now
Most participants did not identify themselves as “enthusiasts” of Rozy, the virtual influencer who fascinates many. Instead, they began following Rozy out of curiosity, as she is the most famous virtual celebrity in South Korea. Participants expressed interest in Rozy and her posts on Instagram because she is virtual, rather than despite being virtual (e.g., participants B and I). This finding somewhat contradicts our initial assumption that Rozy’s followers are enthusiasts who adore her, even though she is a virtual human. According to the participants, they viewed Rozy as an unfamiliar object worthy of observation. Specifically, their curiosity stemmed from a desire to keep up with popular trends and new technologies related to virtual humans, as well as the visual and aesthetic aspects of these digital entities.
Participant J did not know Rozy from Instagram or her daily life. She first learned about Rozy in a class focused on media and virtual humans, which prompted her to begin following Rozy’s Instagram. She then started “studying Rozy as an example of virtual humans for a class assignment.” At that time, Rozy had 50,000 followers, but she now has more than 17 thousand.
For participants who manage multiple Instagram accounts to stay updated on social trends, Rozy represents a crucial trend they must understand to succeed in the media industry after graduation. Our analysis revealed several comments reflecting similar motivations for following and engaging with Rozy’s Instagram account. Consistent with this motivation, some comments treated Rozy primarily as a technological artifact rather than a social actor, either praising or criticizing the virtual human itself, while others focused on the production process by asking about the specific tools used to create her posts (e.g., ‘Can anyone here help me? Which tool does she use?’).
Participant I, who rarely follows celebrities, including actors and singers, is following Rozy not because she admires her as a celebrity, but because she wants to reference the visuals on Rozy’s Instagram feed. Specifically, it is Rozy’s fashion and photography style that have captured her attention. Her appearance feels a bit artificial to me, so I can’t consider her a human celebrity. Whenever I watch her feeds, I think, “Who would create her?” Regarding her writings, I wonder, “There must be some people who wrote this comment.” When I learned about the dark version of Rozy, like a sub-character, I thought, “Who would come up with something like this?” (I) These days, I believe it’s important to stay informed about trends and popular locations, even though I don’t post about them. I use my public account to keep up with social trends. While this information may not be considered knowledge in the traditional sense, I want to gather details about notable exhibitions and newly opened restaurants. (B)
Her virtuality is not always a major drawback
Not all participants viewed Rozy as merely an inanimate object. However, even among those who anticipated human-like interactions with Rozy, we found that they were more focused on the advantages of virtual humans rather than Rozy’s human-like qualities. For instance, a friend of one participant, who had previously admired human celebrities but became disillusioned by their bad behavior, expressed a preference for virtual humans, noting that they cannot engage in negative behaviors like real people do. I recently spoke with a friend who has a strong affinity for virtual humans. She once adored a K-pop idol, but when he was jailed for crimes like drug use and violence, she became disillusioned. She expressed that she grew tired of supporting real people. In contrast, she feels that virtual humans don’t engage in such negative behavior. She said, “Virtual humans feel like they’ll always be there for me. So, I don’t like real people anymore; I only like 2D characters.” (A)
The term “2D people” refers to virtual humans who exist solely in a two-dimensional world, such as animations or cartoons, and do not exist in the three-dimensional world (i.e., the real world). Although it is impossible to meet them in three-dimensional reality, fans feel secure because there is no risk of making mistakes or causing problems, unlike with real human beings. Some individuals, like A’s friend, have learned to stop admiring human celebrities due to their involvement in crimes. While human celebrities can attract fans with their appeal, they can also engage in wrongful behavior.
There have also been instances where Rozy was viewed as an entity with which one could form a social relationship. In these cases, Rozy’s virtual characteristics were seen as advantages rather than obstacles. For example, Participant D appreciated Rozy’s use of familiar verbal expressions that resonate with her generation and her fashion sense. She noted that Rozy, as a 2D entity, never exhausts her in the same way that relationships with real humans can. Rozy communicates in an intriguing way, which draws me closer to her, similar to the man in the movie Her, who experienced a sense of intimacy that went beyond mere admiration for the AI. I believe that in the future, AI and humans could develop relationships that involve sharing emotions rather than just exchanging words. Rozy and other AI could facilitate deeper emotional connections with people. (Do you think you would have such a relationship with Rozy if possible?) Yes. When I engage in conversations with people, I sometimes feel exhausted. Therefore, when I feel lonely and want to talk without the usual fatigue, I think interacting with virtual beings like Rozy could be a good option. I would definitely take advantage of that opportunity. (D)
The interaction involving participant D, as mentioned above, does not refer to commenting or replying on Instagram. Instead, he aimed to have an emotionally engaging conversation with Rozy, who resembles a human and speaks like an emotionally enhanced version of generative AI. Although the virtual human Rozy cannot engage in direct communication with people as a real person would, the participant expressed a desire to communicate and share emotional experiences with virtual humans, akin to AI in the movie. This reflects a subtle hope that the disappointments or fatigue experienced in human relationships could be eased through interactions with virtual beings.
In summary, regardless of whether the participants viewed Rozy as a new technological trend or a flawless celebrity, they all appreciated the qualities of virtual humans that Rozy represented.
Why do we remain mere observers rather than become fans?
The findings presented in the previous sections suggest that Rozy’s followers are not always her “fans,” nor do they necessarily attribute significant meaning to their act of following her. Many participants began following her out of curiosity about new technologies or virtuality, and subsequently chose not to unfollow her, rather than continuing to follow Rozy out of a sense of sustained importance. At the same time, as mentioned earlier, some participants do develop affection for non-human figures and actively engage in fandom practices. In other words, virtuality itself is not the issue. Why, then, do they remain observers rather than become fans of Rozy?
She doesn’t communicate with us in real-time
The participants, who are digital natives, are accustomed to seeing both the polished and glamorous sides of celebrities as well as their more natural aspects through social media, unlike celebrities from the mass media era. It is common for participants to watch celebrities or influencers conduct live broadcasts, interact with followers, and share their activities in real time, which creates a sense of genuine interaction with these public figures.
With the development of various platforms that facilitate direct interactions between celebrities and their fans, Korean fans now enjoy engaging directly with their favorite celebrities, thereby consuming the communication itself (Kang, 2022). Celebrities are no longer distant “stars” in the sky as they were in the past. In this context, Rozy’s followers expect to interact with her directly and in real time. Communicating with Rozy can be a bit odd. While the developers from the company that created Rozy can comment and respond to followers, she is unable to read followers’ comments during a live broadcast in real time. (A) Rozy can't actually go anywhere. For instance, if there’s a very famous bakery, she can post as if she were there, but she can’t physically visit. Additionally, she can’t buy or eat the bread. Because of this, I find that Rozy’s content isn’t particularly practical or informative for me. When Rozy wears pretty clothes, it’s clear she looks good since she is a virtual human with a perfect body and face. I don’t find myself thinking, “Oh, that’s pretty; I want to wear those clothes.” As a result, it’s difficult to feel curious about where she is going or what she is wearing. (B)
Participants expect Instagram influencers to live-stream at any moment, respond to their questions immediately, and share glimpses of their daily lives. However, Rozy does not engage in these activities. For participants accustomed to immediate feedback, Rozy’s method of communication—simply posting images or writing comments—feels slow, one-sided, and frustrating. Additionally, while Rozy’s virtual nature was initially a novelty, it was not compelling enough to compensate for her lack of presence. Instead, it only emphasized her virtuality. I was surprised to see Rozy openly acknowledge that she’s a virtual human. She even discusses it candidly, saying things like, “I can’t really bend my fingers well,” which highlights her status as a virtual being. (E)
Many studies have examined how closely virtual humans resemble real humans in terms of physical appearance. However, participants in this study reported that the virtual humans in current media are already perceived as sufficiently resembling real humans. Mave, a South Korean virtual girl group, performs on stage just like a real idol. They sing and dance with the same flair as human idols, and even their hair movements are perfectly rendered. While watching their performance, I realized that it won't be long before we can no longer distinguish between real humans and virtual ones. I genuinely believe that we may soon struggle to recognize who is virtual. (A)
In summary, while Rozy can gain fame by emphasizing her virtual nature, forming relationships that extend beyond her virtuality is challenging. For Rozy’s followers to become fans, the key factor is not whether she is human or how realistically human-like she appears, but rather the quality and depth of interaction she maintains with her followers. In other words, technological advancement should prioritize creating the feeling of real-time communication and genuine engagement with users, rather than making virtual humans indistinguishable from real ones.
We don’t want to be influenced by Rozy
Digital natives in South Korea often purchase products promoted by human influencers on social media. Participants indicated that the primary reason for buying these items from influencers is their perceived “authenticity.” Having grown up in a digital environment, they have developed a skepticism towards social media content to avoid being deceived. Additionally, they expressed significant resistance to commercials advertised on Instagram. They tend to trust products recommended by ordinary influencers who have personally used them more than those endorsed by celebrities in mass media advertisements.
To minimize exposure to and influence from commercials on social media, they have developed their own strategies. For example, some participants maintain multiple Instagram accounts, separating their personal accounts from public ones and following influencers or celebrities. They want to stay informed about these celebrities but do not wish to see their advertisements. Furthermore, they utilize the platform’s tools, such as hashtags, to avoid unwanted ads. For instance, participant B mentioned that she uses hashtags on Instagram to prevent encountering ads inserted between her feeds. I used to search extensively on Naver, but it feels like it’s mostly ads, making it hard to trust the results. It seems that everything at the top of the search results is an advertisement. That’s why I’ve started using Instagram to search with hashtags. For example, when I want to visit Seongsu-dong, I search for “#SeongsudongRestaurants” on Instagram, where I find more authentic photos. While there are images posted by influencers, they are usually marked as “ad” or “sponsored.” By considering these labels and filtering the photos accordingly, the results feel more trustworthy, so I’ve been using hashtags a lot lately. (B)
Although participants believe that promotions from influencers who demonstrate authenticity are more trustworthy than traditional commercials, they still employ their own verification methods before purchasing products endorsed by influencers. For instance, they closely examine whether an influencer genuinely uses a product and whether the product’s company is reputable. Additionally, some participants who hold negative views toward commercials are hesitant to follow influencers, as these influencers are often paid to promote products. This indicates that even when participants buy products endorsed by social media influencers, they strive to be discerning consumers who avoid being commercially exploited or misled when endorsements do not meet their verification standards.
In this context, Rozy, as a virtual human, fails to meet these standards because she lacks real experiences with products and cannot convey personal thoughts or feelings based on such experiences. Unlike human influencers who engage in live broadcasts and eat products they are promoting, making followers feel connected to the experience, Rozy cannot actually use or purchase the products she advertises. As a result, followers are unlikely to trust Rozy’s promotions. If Rozy posted fun videos like a human influencer and engaged in real-time conversations, I would feel more attached to her. People might think, “Oh, she has this one” or “This looks pretty.” However, Rozy only shares photos of the products she’s advertising. While they are visually appealing, they don’t inspire me to buy the products; instead, I feel a sense of rejection. (H) The reason I don’t purchase products from Rozy Promotions is primarily that she’s not a real person. Rozy cannot genuinely use any product. In contrast, human influencers, even if they lack expertise in cosmetics, can at least demonstrate the products to their followers or explain their benefits. However, Rozy can only be shown holding the products. (A)
The participants acknowledged that virtual humans like Rozy can generate buzz as product models, which can be beneficial for companies employing them. However, the lack of a virtual human presence results in a loss of authenticity for digital natives, who are used to purchasing products from human influencers known for their authenticity and trustworthiness.
Something 2D animation characters do, but Rozy doesn’t
Participants who couldn’t be fans of Rozy due to her lack of presence and authenticity noted that they do enjoy virtual characters from animations, even though these characters are not real humans either. For instance, posts from the official Instagram account of the Korean virtual boy band Plave often attracted comments that resembled those directed at human celebrities. In one post featuring Noa standing on a beach, users responded with remarks such as ‘Where is this prince from? Unbelievable!’ and ‘Coca-Cola advertisers should check out this photo.’ Rather than discussing the technologies behind the character or questioning how the image was produced, commenters focused on Noa’s appearance and appeal. What’s the difference?
According to the participants, the major difference between Rozy and the characters they love is the presence of “complex character.” They stated that it doesn’t matter whether they love a real human, like a K-pop idol, or a virtual character from animations or cartoons. To form a stable and sustainable relationship filled with affection, a character must possess complexity and depth, supported by rich narratives—qualities that Rozy does not exhibit on Instagram. Regarding characters from animations, there is a concept called “character interpretation,” where people create hypothetical stories, such as “if this character were in a certain situation, they would probably do this or that.” However, we cannot create such hypothetical stories about Rozy. Rozy is simply a new technology, and there isn't much to discuss. (C) The lion character from Kakao has a backstory that explains why it is very upset about not having a mane. This backstory makes the character more attractive and interesting to people. Similarly, if Rozy had a compelling backstory, it might resonate more with the audience. Such emotional connections can sustain interest and enhance a character’s appeal. However, in Rozy’s case, people simply find her “interesting” and then move on. (E)
Although characters from animations are not real humans like Rozy, they possess complex personalities and engaging stories. Additionally, fans have memories associated with watching these animations over time. The issue is not merely that Rozy is a virtual human; rather, it is that there is nothing to imagine, discuss, or interpret, which hinders followers from forming a closer connection with Rozy. I really like Shin-chan. He has his own episodes that are cute and nostalgic, which Rozy lacks. If there were ongoing animations featuring Rozy as the main character, I might be more interested. However, since she is just a virtual influencer, there isn’t much to engage with. (F)
Additionally, virtual humans like Rozy are developed using advanced technology, which restricts fans’ ability to create derivative works. For instance, fans can produce new animations or fan-made videos of 2D characters and comic figures, as well as edit videos featuring real human idols. However, this is not possible with Rozy. Isn’t the content available for fans of Rozy quite limited? Fans cannot create content about Rozy; only the developers can do that. There aren’t multiple videos that fans can edit to highlight key moments or create compilations with continuity. As a result, she’s perceived merely as a hot influencer, and that’s about it. (J)
Interpreting a character goes beyond mere observation; it involves the joy of engaging with the characters as active participants and creating additional works. However, technical constraints can hinder these fan activities, limiting the development of affection among fans.
Discussion
This study explored the characteristics of followers of virtual humans, their motivations for following, and their perceptions of virtual influencers compared to human celebrities. The aim was to understand the current role of virtual humans in the entertainment industry and how this role contrasts with that of human influencers and celebrities. By conducting online observations and in-depth interviews with Rozy’s followers, we discovered that these individuals cannot be simply categorized as fans. While some followers were virtual accounts, companies, or unidentified users, the digital natives in the study followed Rozy out of curiosity about technology and popular trends rather than genuine affection. Additionally, Rozy’s lack of real-time interaction, authenticity, and narrative depth hindered participants from forming a connection with her or viewing her as trustworthy. Ultimately, Rozy’s followers function more as observers than fans or consumers, challenging assumptions made in previous studies (e.g., Gerlich, 2023; Liu et al., 2025). In essence, the followers do not regard Rozy as a traditional celebrity, as their interactions with her differ fundamentally from the typical celebrity-fan relationship.
These results challenge our initial assumptions regarding Rozy’s followers—not only that they are necessarily her fans, but also that her virtual nature would hinder the formation of such a fandom. These assumptions have been common in studies of virtual influencers, which often regard physical human attractiveness as a key factor in shaping relationships between celebrities and fans (e.g., Choi and Rifon, 2007; Onu et al., 2019; Santoki and Agrawal, 2025). However, the findings suggest that it is the audiences’ interpretive practices and interactions with virtual characters, rather than their physical realism, that are vital for developing fandom and deep emotional connections. This aligns with previous research indicating that audiences seek a promise of “authenticity” through consistent and intimate interaction (Abidin, 2015; Marwick, 2013).
In the initial stage of this study, we treated follower counts and visible engagement metrics, such as likes and comments, as reasonable indicators of popularity and, by extension, fandom-like attachment. This approach aligns with a broader industry and media tendency to interpret platform metrics as social proof, a shortcut particularly relevant in the Korean context. In South Korea, Rozy was frequently portrayed in mainstream media as a “celebrity” with a large Generation Z fanbase and significant public enthusiasm. Her Instagram following and brand collaborations were often cited as evidence of her cultural impact. Moreover, much of the literature on virtual influencers, especially in marketing and effects research, implicitly models followers as potential consumers or fans, viewing following as a sign of relational commitment.
However, our analysis revealed something striking: followers are not always fans. Rozy’s case highlighted how the conditions of virtual influence can undermine the assumption that follower counts equate to genuine fandom. The follower ecology included a significant number of non-human, corporate, and unidentifiable accounts, while the dominant interaction style in comments was often non-relational—characterized by formulaic, low-meaning responses or technical evaluations. These patterns complicate the media’s implied equation of “many followers” with “many fans,” prompting us to closely examine how followers actually position Rozy and the types of audience relationships this positioning facilitates or limits.
A key finding of this study is that participants did not view Rozy’s virtuality as a weakness or disadvantage. In fact, it was this virtuality that attracted them to Rozy in the first place. This aligns with previous research suggesting that Rozy could gain fame by embracing her virtual identity rather than obscuring it. For instance, Kang and her colleagues (2023) noted that Rozy openly acknowledges her virtual nature with statements like, “I never age” and “I don’t have a national ID card because I’m not a real human.” Similar to how human celebrities share behind-the-scenes moments to showcase their authentic selves, Rozy reveals her “virtual human” identity.
However, Rozy’s strategy of openly embracing her virtuality often distances her from her followers, rather than creating a sense of authenticity. Unlike human influencers who share their unfiltered lives (AlRabiah et al., 2022; Wang and Taylor, 2025), Rozy presents a polished image more akin to traditional celebrities. Participants noted this distinction, expressing that they found human influencers more relatable. While Rozy fits the definition of a social media influencer as described by Piehler et al. (2022), her engagement style resembles that of traditional celebrities, fostering admiration rather than intimacy with her audience (Turner, 2014). As a result, Rozy’s virtuality may hinder emotional connection, positioning her more as a spectacle of technological achievement than as an authentic social figure.
This study has several limitations. First, it focused on a single virtual human, Rozy, which may limit the representativeness of the findings. While Rozy is recognized in South Korea as a leading “virtual influencer” and is one of the most technically advanced virtual humans featured in mainstream media and advertising, she may not reflect the broader category of virtual humans. Audience expectations, follower dynamics, and interaction norms can vary significantly across different genres, such as virtual idols, VTubers, and game characters, as well as across various production models. Consequently, as new and differently styled virtual humans emerge, the forms of audience attachment observed in this study may also evolve. Second, the participant demographic was predominantly women in their 20s, which resulted in limited diversity in age and gender. Additionally, since we interviewed only a subset of Rozy’s followers rather than her entire follower base, generalizing the characteristics and motivations identified in this study is challenging.
Third, given the rapid evolution of virtual human technologies and platform practices, our findings should be considered time-bound to the current stage of development. Longitudinal research is necessary to track how advancements in realism, interactivity, and content production influence follower interpretations and relationships over time.
While many studies have explored audience perceptions of virtual influencers, particularly in marketing contexts, this study delves into the less examined intersection of virtual humans and celebrity studies. First, it introduces a novel type of relationship between virtual humans and human audiences. Rather than relying on conventional categories like fans, consumers, or users, this study categorizes Rozy’s followers as “observers,” providing a valuable theoretical framework for understanding the complex behaviors of audiences in the digital media landscape. Second, this research contributes to the field of celebrity studies by emphasizing the unique position of virtual influencers, distinguishing them from traditional categories such as celebrities, stars, ordinary influencers, or two-dimensional characters. Third, our findings demonstrate that physical embodiment is not a prerequisite for achieving celebrity status; instead, realistic, real-time interactions and compelling storytelling are essential for building a fan base. Furthermore, this study identifies authenticity as a key motivation behind followers’ preferences for ordinary influencers and their subsequent purchasing decisions—an important trait of digital natives who are often skeptical of overtly manufactured promotional content. Followers can become genuine fans—whether the influencer is human or virtual—when the influencer engages in active communication and credibly shares personal experiences (Lim and Lee, 2023). Accordingly, participants, predominantly digital natives in their twenties, see Rozy’s lack of real-time interactions—like live Instagram broadcasts, prompt responses to comments, and sharing of everyday experiences—as a major drawback. These findings indicate that future advancements in virtual human technologies should focus on improving interactivity and immediacy instead of just aiming for visual realism. Additionally, the study revealed that storytelling and a character’s multifaceted appeal are more important for fostering fandom than whether a celebrity is virtual or real.
Footnotes
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported in part by the Special Research Grant of the Korean Association for Broadcasting & Telecommunication Studies and GS Retail in 2022. This research was supported by Hallym University Research Fund, 2026(HRFN-202603-005).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.