From Novelty to Credibility? A Longitudinal Structural Topic Model of Audience Discourse towards Lil Miquela on YouTube
This longitudinal study of YouTube comments reveals that audience discourse toward the virtual influencer Lil Miquela did not follow a uniform trajectory from novelty to credibility, but instead exhibited selective normalisation characterized by declining identity confusion, rising creative engagement, and significant fluctuations driven by content gaps and sponsorship contexts.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the digital world, a new kind of celebrity has emerged: the virtual influencer. These are not human beings with cameras and microphones, but fully computer-generated characters who post photos, sing songs, and interact with fans just like real people. They are managed by teams of designers and marketers, yet they are presented to the public as if they have their own lives, feelings, and personalities. For years, researchers have wondered how people react to these synthetic figures. Do we treat them as strange novelties at first, and then slowly grow to trust them as they age? Or does the awareness that they are fake always keep us at a distance? This question sits at the intersection of psychology and marketing, exploring how we decide what is real and what is trustworthy in an age where the line between the two is increasingly blurred.
A team of researchers set out to answer this by watching how people talked about one of the most famous virtual influencers, a character named Lil Miquela, over a period of five and a half years. Instead of asking people to fill out surveys or testing them in a lab, the team collected thousands of public comments from YouTube videos featuring the character. They looked at everything from the very first questions about whether she was real, to later discussions about her music, her fashion collaborations, and the brands she promoted. By using a method that groups similar comments together to find hidden themes, they mapped out how the conversation changed from 2021 to 2026. They wanted to see if the audience eventually stopped asking "Is she human?" and started treating her like a normal celebrity, or if the conversation took a different path entirely.
The results revealed that the story of how we accept virtual influencers is not a simple straight line from confusion to trust. In the beginning, a significant portion of the conversation was indeed focused on identity. People were genuinely puzzled, asking if Lil Miquela was a robot, a human, or something in between. They questioned whether her interactions were scripted or if she had real feelings. However, as time passed, these questions about her nature did not simply vanish to be replaced by blind trust. Instead, the focus of the conversation shifted dramatically. The number of comments asking about her identity dropped by more than six percentage points, but this did not mean people stopped caring about her reality. Rather, the conversation moved toward what she actually did.
The most striking change was in the type of content people discussed. As the years went on, comments about her music and creative work surged, rising by nearly fourteen percentage points. At the same time, a massive wave of attention appeared whenever the character returned after a long break or released something new. This "return and renewed attention" topic grew by over twenty-three percentage points, becoming the most dominant theme in the later years. This suggests that the audience's engagement was not driven by a gradual maturation of trust, but by the rhythm of the content itself. When the character was active and releasing songs, people talked about the art. When she was silent, they talked about her absence. The idea that people simply "got used to" her and moved on to a steady state of credibility was not supported by the data.
The study also looked closely at how the audience reacted when the character was promoting a product, such as a car or a computer, versus when she was just sharing personal updates. When the videos were clearly sponsored, the conversation changed in specific ways. People talked more about the products themselves and the technology behind the advertising. They also discussed the corporate control behind the character more frequently. Interestingly, while one might expect that a commercial pitch would make people more skeptical, the data showed a complex picture. In sponsored videos, the usual questions about whether the character was "cringe" or socially acceptable actually decreased. This did not mean people trusted her more; it meant the topic of conversation had been redirected entirely toward the product and the technology, pushing other concerns aside.
Crucially, the researchers found that who posted the video mattered just as much as whether it was an advertisement. When a brand posted a video on its own channel, the audience was different from when the character posted on her own channel. In some cases, the apparent difference between sponsored and non-sponsored comments disappeared once the researchers accounted for who had uploaded the video. This suggests that the environment of the channel shapes the conversation as much as the commercial intent does. A product-focused audience on a car manufacturer's channel is not the same as the fan community on the influencer's own page, and they bring different expectations to the discussion.
Ultimately, the study paints a picture of a relationship that is selective and event-driven rather than a smooth journey toward acceptance. The audience did not uniformly decide that the virtual influencer was credible. Instead, they engaged with her based on a mix of factors: the novelty of her identity faded, but her creative output and the timing of her appearances took center stage. The conversation was punctuated by specific events, like new music releases or long periods of silence, which dictated what people talked about more than any underlying shift in their beliefs. The findings suggest that for virtual influencers, maintaining an audience is less about proving they are real and more about managing the flow of content and the context in which it appears. The trust that develops is not a single, solid state, but a collection of different reactions that change depending on what the character is doing and where the audience is watching.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.