Virtual influencers:
How AI avatars are revolutionizing advertising

AI-generated representations of people are causing debate. Deepfakes impressively demonstrate how easily artificial intelligence can imitate and deceive reality. The viral image of Pope Francis in a fashionable down coat or the deceptively real video of Olaf Scholz announcing a ban on the AfD illustrate both the fascination and the risks of such technologies. But what happens when representations bear no relation to reality? What if digital avatars are created artificially from scratch? This is precisely where virtual influencers come in, redefining not only marketing but also our relationship with technology.

Spike Jonze’s 2013 film “Her” seems more relevant today than ever. It centers on Theodore, a lonely man who develops a relationship with “Samantha” – an AI with an empathetic voice, spoken by Scarlett Johansson. Samantha simulates emotions so convincingly that she becomes Theodore’s confidante, blurring digital and real worlds.

This fictional story is becoming more and more tangible in reality. OpenAI’s voice “Sky” was criticized because it was strongly reminiscent of Scarlett Johansson. Although the company emphasized that no imitation was intended, the case shows how closely AI today operates on the border between reality and imitation.

An episode of the British science fiction series Black Mirror “Be Right Back”, also from 2013, also seems prophetic today. In it, a grieving woman creates a replica of her deceased partner with the help of an AI. What initially provides comfort quickly becomes a source of unease – the perfect simulation appears lifeless and incomplete.

This scenario has long been a reality in China, where companies offer avatars of the deceased that are created using photos and videos. One family used this technology to conceal the death of their son from a 93-year-old mother. However, critics warn that such practices disrupt the grieving process and raise ethical questions about consent and responsibility.

 

Lil Miquela: The pioneer of virtual influencers

Lil Miquela, one of the best-known virtual influencers, on the other hand, is a completely imaginary avatar with no real-life role model and has had an impressive career since her Instagram debut in 2016. She presents herself as a 19-year-old half-Brazilian living in Los Angeles and now has around 2.5 million followers on Instagram. Her feed is a mix of fashion, art and music that comes across as so authentic that many of her fans didn’t know she wasn’t a real person at first.

Miquela is backed by the technology start-up Brud, which develops and controls her character. With a team of designers, authors and technicians, Miquela is not only staged as a fashion icon, but also as a singer. Her music, which she has been releasing since 2017, moves stylistically between house and pop. Songs such as “Not Mine” and “Hate Me” generate millions of streams and show that AI influencers can reach an audience of their own.

Miquela is not just a product, but also a personality. She appears in interviews, tells personal stories and interacts with her fans – all carefully planned by her team. Her success has paved the way for many more virtual influencers, proving that digital personalities are able to simulate human connections and represent brands effectively.

Anna Indiana: The first fully AI-generated singer

While Lil Miquela is controlled by a human team, Anna Indiana is an experiment in complete autonomy. She is being promoted as the world’s first singer-songwriter created entirely by AI. Her debut song “Betrayed by This Town” was generated using tools such as ChatGPT and Musicfy. The melody, lyrics, Anna’s appearance and voice are all the result of AI algorithms. However, she currently only has 75 monthly listeners on Spotify. “Miles of Smiles”, a video in which she explains in advance how she and the song were created, has been viewed 17,000 times on YouTube. (as of November 2024)

Anna’s performances are met with mixed reactions. Critics find her performance “soulless” and technically immature. Her visuals are part of the “Uncanny Valley”, a concept from robotics and computer graphics that describes how human-like representations are perceived. The closer a figure comes to a human appearance, the more positive the reaction is at first. However, if it appears almost, but not quite perfect, a feeling of unease arises – triggered by subtle deviations such as rigid facial features or unnatural movements. It was precisely these inconsistencies, combined with Anna’s lack of emotional expression and stiff movements, that made acceptance even more difficult.

Nevertheless, Anna Indiana’s existence shows what is possible with AI in music. Proponents argue that the technology is still in its infancy and will improve over time. Opponents, on the other hand, question whether an AI artist who only imitates human creativity deserves a place in the music world at all. Her singing and vocal performance, which can be heard on platforms such as Spotify and YouTube, is for everyone to judge for themselves.

 

Emma: A missed opportunity for the German National Tourist Board

The German National Tourist Board (GNTB) recently caused a stir with its virtual influencer Emma – but not in a good way. Emma was developed to promote Germany as a travel destination and was intended to be a modern, digital ambassador. But instead of enthusiasm, Emma reaped a shitstorm. The use of AI avatars in advertising poses many challenges, as this case shows.

There were numerous points of criticism: her performance seemed technically immature, with typical AI errors such as unnatural movements and blurred details. In addition, Emma appeared stereotypical as a blonde woman with model measurements and was unable to convey an authentic travel experience. The lack of authenticity in particular was perceived as a major problem by many users. The associated AI chatbot also provided inadequate and incorrect information, which further diminished the overall impression.

In addition, the question arises as to whether the topic of tourism marketing, in which people are to be inspired to have real experiences – i.e. travel – is not simply unsuitable for artificially created realities.

Emma’s case shows that technological innovation alone is not enough. Virtual influencers must be carefully planned and mature in their presentation in order to be accepted by the target groups. Otherwise, brands risk losing their credibility and creating negative associations.

Trigema: The iconic monkey Charly as an AI fashion influencer

Trigema has ventured into the digital era with its iconic monkey Charly and shows how tradition and innovation can be successfully combined. As an AI fashion influencer, Charly not only presents fashion, but also comments on current affairs – all in a completely digital, AI-generated form.

The campaign uses generative AI to make Charly more flexible and versatile. Unlike previous versions, which relied on classic animation techniques, the AI technology allows Charly to appear in completely new contexts and develop more personality. He can act in social media formats or TV commercials while authentically representing the company’s values.

The balance between the company’s long tradition and the modern approach of the campaign is particularly noteworthy. Bonita Grupp, who took over the management together with Wolfgang Grupp Jr., emphasized that it was important for Trigema to retain both the Charly monkey and the company values. The further development of Charly should not only show the modern orientation of the brand, but also reflect its roots in sustainability and family values.

The use of AI also gives Trigema creative freedom and enormous flexibility in terms of presentation. Charly can react dynamically to current trends and topics, making him a future-proof advertising medium. The campaign sets a new standard for the use of virtual influencers in marketing.

 

How consumers see virtual influencers

Current studies show that consumers have ambivalent perceptions of digital avatars. On the one hand, their efficiency and almost unlimited creative possibilities are appreciated. AI avatars in social media marketing can be customized, adapted at any time and used without the restrictions of real people. They enable cost-efficient productions and offer brands the flexibility to tailor their content precisely to target groups.

On the other hand, many consumers express concerns. Unrealistic ideals of beauty, which could be promoted by perfect, symmetrical faces and flawless images, are particularly criticized. This perfection is perceived as problematic, especially among younger target groups, as it could have a negative impact on their self-image and promote eating disorders or mental stress. Gen Z in particular is open to AI avatars, but expects authenticity and diversity. The apparent lack of authenticity of virtual characters is another point of criticism, as many users miss an emotional connection to real people.

Interestingly, many people do not recognize any difference between edited images and AI-generated avatars – as long as the representation is realistic. This shows that the acceptance of virtual influencers depends heavily on their appearance and impact. Perfect faces often appear unnatural and meet with rejection, while natural imperfections can increase credibility.

Especially in sensitive areas such as cosmetics or health, most consumers prefer real people. Here, they expect authenticity in order to gain trust in the products. At the same time, however, there are areas such as technology or fashion where virtual influencers are perceived as modern and innovative, which underlines their potential in these segments.

This ambivalent attitude makes it clear that the success of virtual influencers depends heavily on how carefully they are designed and the context in which they are used. Unthinking use can easily lead to mistrust, while a thoughtful approach can create an innovative and successful brand experience.

 

Tips for the use of AI avatars

The success of AI avatars and virtual influencers depends on well thought-out planning, precise implementation and target group orientation. The following recommendations are designed to help brands make the most of the potential of this technology and avoid pitfalls.

 

Do’s:

1. Realism and authenticity:
   AI models should appear as realistic and authentic as possible. It is essential that they have a personality with a comprehensible background story and details that the target group can identify with.
2. Consider diversity:
   A variety of ages, body shapes and skin colors increases the acceptance and appeal to different target groups. This is particularly important in areas where brands represent values such as inclusion and diversity.
3. Context-related use:
   AI avatars are suitable for specific tasks such as the presentation of products, landscapes or accessories, as long as they do not raise false expectations. AI content can therefore be a modern and efficient approach for technical products.
4. Preliminary tests in closed groups:
   To avoid possible negative reactions, avatars should be tested in user groups. These tests provide valuable insights into the effect of appearance, language and behavior.
5. Target group-specific positioning:
   Before using AI, it should be checked whether the target group is open to AI avatars. For topics with a high emotional or personal connection (e.g. health, charity), real influencers are often the better choice.

Don’ts:

1. Avoid perfectionism:
   Perfectly symmetrical faces and flawless portrayals can quickly appear unrealistic or uncanny. AI avatars should have natural imperfections in order to appear more credible.
2. Refrain from making exaggerated promises:
   Products such as cosmetics or food should show realistic results. Exaggerated representations by AI can be perceived as misleading and damage trust in the brand in the long term.
3. Avoid early publication:
   Immature AI models or chatbots, as in the case of “Emma”, can cause considerable damage to the brand. AI applications should only be published after extensive testing and optimization.
4. Unclear labeling:
   AI-generated content must be labeled transparently.
   This protects against legal problems and increases the trust of the target group.
5. Overestimate the first mover effect:
   Simply using an AI avatar without offering any real value is risky. Brands should ensure that AI content creates added value, for example through personalization or innovative interactions.

 

AI avatars should not only be visually appealing, but also functionally convincing and create real added value. Brands must use the technology in such a way that it not only attracts attention, but is also perceived positively in the long term. In this way, brands can take advantage of AI avatars without jeopardizing credibility and consumer satisfaction. The right use creates innovation and strengthens the brand image in the long term.

 

Virtual influencers: opportunity and responsibility

Virtual influencers are far more than just a passing marketing trend – they embody the combination of technological innovation and social change. AI-powered social media campaigns with avatars open up unprecedented opportunities for brands to reach new target groups, tell creative stories and position themselves as pioneers of digital technologies.

However, this opportunity also comes with responsibility. Success depends on how credible and authentic these digital figures appear and how ethically they are integrated into campaigns. Only those who harmonize authenticity and technology can gain the trust and enthusiasm of consumers in the long term. Virtual influencers are not just a gimmick, but a strategic tool for the future of marketing.