With tousled pink hair, feline eyes and a prominent decolletage, Aitana Lopez, 25, is the kind of model who could sell anything. Geordies would queue up to buy coal from her. Her Instagram account, which has nearly 200,000 followers, shows her in outfits from brands including Victoria’s Secret, Brandy Melville and Guess. For her modelling work, she reportedly makes up to £9,000 a month. On Fanvue, a creator platform, devotees can pay for private content and exchange messages with her.

“Ready for the weekend in my leather goddess outfit,” she announces in one recent caption. “What word do you think best defines my personality or style?”

How about ‘invented’? The best thing about López, from her bosses’ point of view, is that she does not exist. Forget not waking up for less than $10,000 a day, López does not need to be paid at all. She works all hours of the day and night, never takes holidays or falls ill or complains. She is just pixels and imagination.

Her creators, Spanish company The Clueless, conjured her as a solution to the inefficiencies of dealing with human models. “We started analysing how we were working and realised that many projects were being put on hold or cancelled due to problems beyond our control,” said Rubén Cruz, the agency founder, in an interview with Euronews. “Often it was the fault of the influencer or model and not due to design issues.” The logic is clear: why deal with a temperamental reality when you can rely on a perfect image?

“We did it so that we could make a better living and not be dependent on other people who have egos, who have manias, or who just want to make a lot of money by posing,” explains Cruz.  

López is only one of thousands of AI influencers and models who are rapidly becoming big business. Digital characters such as López have been around for a while, since the cartoonish, pixelated Lil Miquela [one of the first AI models, who has nearly three million Instagram followers], but the new iteration is infinitely more polished. What’s also novel  – and more important – is the extent of the monetisation.

“AI creators made up 15 per cent of our revenue last month which is really impressive,” says Will Monange, the co-founder of Fanvue. “That’s up more than 100 per cent on the month before, and that’s a progressive increase. Interest is growing.”

For Monange, López has been successful because she has a talented team behind her. Given that the cost of creating an artificial persona is almost zero, all the skill is creating a figure who stands out from the crowd.

“It can’t be generalised as ‘it’s easy to do,’” he says. “There is genuine artistic talent in making the AI character believable. Lots try, but those who get that level of quality where you can’t tell the difference are talented, creative people who are choosing to build out this form of art. Then you have to do the right things on social media to grow their following. There are lots of good-looking [real] people in the world, but it doesn’t mean they go viral.”

For now, most of the social media following of these creators is on Instagram, where still images continue to hold some sway, as opposed to TikTok, which is pure video. While still images of artificial creators are uncannily lifelike, videos are not at the same level of fidelity to reality. But they are improving constantly; it is a matter of time before you will not be able to tell the difference.

He adds that the flesh-and-blood creators should not be over-alarmed by the rise of these AI competitors. “It’s like the broader fear with AI,” he says. “There’s a lot of pessimism but fundamentally it’s a tool for creators. We’re talking about a creative person who generates an audience. At the moment they have to be the face of that creation. What this is doing is unlocking the possibility of people who can unlock an audience but don’t necessarily want to be the face of that creation. They can use this tool to build an audience and connect with fans. It’s opening the door to creative people. I don’t see it replacing anybody.”  

Shahnaz Ahmed, director of creative and innovation at The Social Element, a social media agency, says AI influencers represent a turning point in marketing and advertising. “Being digitally programmed, AI creators can be the most accessible version of influence we’ve seen to date,” she says.

“They can speak any language we want, create content quickly, stay constant and be fully adaptable. Research has also shown that human beings find it easier to open up more to a virtual person than a real one: think ‘confession box’, a safe space to be honest with no judgment. It has the possibility to support our mental health and wellbeing needs in ways humans can’t.”

Not everyone is convinced we are on the right path. Cammy Crolic is an associate professor of marketing at the University of Oxford, who is researching AI models and influencers. “From a brand perspective, there’s a lot of safety in working with virtual influencers, because they don’t come with any reputational concerns,” she says. “There’s a lot more control over what the message says and how they present themselves.” In other words Aitana López will not be photographed falling out of a nightclub, or say something controversial in an interview.

“But from a consumer’s perspective, the research does seem to coalesce around the idea that hyper-realistic influencers connect better with consumers than more cartoonish ones,” says Crolic. “But there is another side of the coin, which is that people are aware that this takes away business from real people.”

In particular, she says, there is a risk that AI diversity replaces the real thing. Rather than hire models from ethnic minorities, AI companies can simply generate a diverse artificial influencer. Earlier this year, Levi’s faced a backlash after it announced it was going to start using AI models to improve “inclusivity”. 

“It’s very easy to simulate diversity but this is taking away important work from people who have historically not had as much representation in the media,” says Crolic.

Her research is unfinished, but so far suggests the flawless AI beauties like López also reinforce unhelpful ideals. “Tentatively, what we’re finding is that it highlights a feeling of objectivation, which negatively impacts their self-esteem. The media objectifies women, they’re now creating perfect bodies. Society must want women to be pretty things to look at as opposed to real, nuanced women with backgrounds and history and flaws that make them interesting. [The rise of AI influencers] highlights that it’s all about physicality.”

Crolic says that in the time she has been focussing on this topic, her students have become much more aware of the presence of AI models and influencers in advertising and marketing. “When I started introducing this to my classes a couple of years ago I would show them pictures of Lil Miquela and they would be amazed she wasn’t real,” she notes. “There has been a big uptick in awareness of AI influencers’ use, which brands are using them and the money generated around it. All these things make it more important to study this.”

Another open question is the psychology of the people ‘interacting’ with these personalities online. Creating an advanced mannequin for a pair of jeans or a dress is one thing, but up until now, part of the appeal of sites such as Patreon, OnlyFans or Fanvue has been the idea that you are interacting with a real person. Often it is a man interacting with an attractive woman. But as Monange points out, when these exchanges are mediated by a digital platform, there is no way of knowing that you are getting the real deal, a chatbot or a middle-aged man pretending to be a pink-haired, Spanish 25-year-old.

“Ultimately any social-media based interaction is a parasocial relationship. The fan feels connected but they’ve not met the person,” says Monange. “I think different generations will have different views. As people grow up with it as more of a norm, it will feel completely normal that someone is a digital creator. I think [AI influencers] are going to be a significant part of a multi-billion dollar economy.”

Others believe it will be some time before a virtual influencer can replicate all the benefits of a real one. “Creators do more than encourage clicks and likes, they connect deeply with those they are talking to” says Ben Jeffries, of Influencer.com. “Where a virtual influencer might have short-term success, it would likely take a long time to build that long-term trust and emotional connection. There will also be legitimate questions about regulation and safety.” Then there is the issue of cost: while you might not have to pay a character, paying the staff involved in managing and creating one probably costs more than using a human.

“AI influencers can’t replicate human experiences,” says Jago Sherman, head of strategy at the Goat Agency. “They can’t be a parent, can’t feel emotions, can’t have real relationships, and they can’t talk about experiences like using a skincare product or loving a particular food. They’re more like a billboard or digital ad, and the reason influencer marketing is so powerful is that consumers want to buy products from real people that they can relate to.”

Yet it is undoubtable that as realistic as the pictures of Aitana López are becoming, they are cartoons compared to what will be possible in the future. Before long there will be videos indistinguishable from reality. Fashion has always been about aspirational images, even illusions: now the models are made up, too.