Fake Smiles & Real Money: The World of Virtual Influencers & AI Celebs
What are the implications of letting a perfectly manufactured image do all the work?
I’m a big fan of Charlie Brooker’s dark sci-fi anthology Black Mirror (I even wrote an in-depth review of the fourth season). The show has been around since 2011, eventually finding its home on Netflix, and captivating audiences with tales of technology and humanity gone awry, and asking the question “What if?” Another talking point of the show — and a possible reason why it became so popular — is that most of the technology showcased isn’t too far in the future, especially at the rate we’re advancing. It was even confirmed that one of the episodes, “Metalhead,” was inspired by the robot AI “dogs” called BigDog, from Uncanny Valley connoisseurs (and tech/robotics company) Boston Dynamics.
Over the years, the show has gained a large fanbase, enough to the point that our lexicon can now contain the phrase, “What episode of Black Mirror is this?”, in response to something that seems so eldritch and dystopian that it can’t possibly happen in real life (despite the warning that it very well may). And more often than not, the truth is we’re living right in the midst of one.
Part I: The Rise (and Flux) of Influencer Marketing
Unique ways to advertise to the consumer aren’t anything new. Celebrities of the traditional variety — usually actors and singers — become the faces of brands and populate commercials and advertisements galore. But, the advent of social media allows regular people to develop notoriety overnight or overtime, filling niches and becoming so-called experts with so-called trusted opinions. It was only natural that companies took note and decided to adapt.
Many companies themselves are born on Instagram and Facebook through extensive advertising — no longer do you need a brick-and-mortar store and traditional PR to drum up business. With the nuance of parasocial relationships built through social stars (they feel like your neighbor, your friend, someone you trust) and the ease of starting (you can do it too!), it’s no wonder influencing became a legitimate career. As of writing this, there are 13 million posts alone that are tagged with #ad and over 3.5 million with #sponsored. On top of the adverts, there’s significant money to be made per post for some of the largest influencers and celebs — in the thousands.
Agencies were created specifically for managing influencers and influencer marketing, a field that was not around a mere ten years ago. And because of the voracious appetite of social media-savvy millennials, it’s being looked at as a potential job opportunity — if you hit the media (and genetic) jackpot, you could be set for life.
A poster needs to strike the precise balance between too many sponsored posts or mismatched product pushes (which can be perceived as “selling out”) and regular, non-sponsored postings. A typical user trusts an influencer to deliver real reviews and products that they genuinely like, but it’s hard to determine if — and how much — the presence of a paycheck has any influence. That’s not to undermine how hard an influencer works to generate that relationship with their audience, though — how else would they be able to, well, influence?
Still, many marketing initiatives are starting to veer away from those who have too many followers, like Kylie Jenner, who sits at a whopping nearly 200 million. Sure, she’ll still get tons of sponsorships and make a hefty amount per post, but brands are looking at tapping into other key influencer types outside of the highest celebrity tier, like micro-influencers and nano influencers, who’d have far fewer followers than Jenner but enough of a following to generate some influence. Plus, it’s significantly cheaper on a brand’s end to go the small creator route, and they can achieve a key component of success ROI much easier: authenticity and trust. People want to believe that the influencer is using the product, whatever it may be (and if they aren’t, it’s actually a violation of the FTC).
And of course, influencer marketing can hit a low when the partnership goes awry, or the aforementioned authenticity is questioned. Who can forget the suitably infamous Fyre Festival from 2017 — whereupon celebrities and influencers like Kendall Jenner promoted the festival on Instagram for $250,000 a post (without initial disclosure), only to have the entire thing result in chaos, wire fraud, and imprisonment.
Sure, this is by far an extreme example of marketing (and pretty much everything you could imagine) gone wrong, but it leads us to ponder a way to avoid such atrocities and what may come. Despite the lows, it seems that this type of marketing is here to stay, and will evolve with the times. However, that evolution posits a question — with technology’s looming omnipresence constantly advancing, what will its intersection be on social media?
Part II: Holographic Celebrities & AI Influencers & Fake People, Oh My!
One of the later episodes of Black Mirror, “Rachel, Jack, and Ashley Too,” employs Miley Cyrus in one of the lead roles as a shiny pop sensation, performing robotically and clearly hating her job. (Whether this is a commentary on her time as Hannah Montana is unknown, but I digress.) When she falls into a coma at the hands of her manager/aunt, she seeks to keep her that way, since she can easily replicate her singing voice into a computer-generated persona, and not worry about voicing any distaste.
This isn’t too far from the truth, unsurprisingly. Hologram performances by Tupac in 2012 Coachella and Michael Jackson at the Billboard Music Awards in 2014 shocked the world, and simultaneously treated as magnificent feats of technology while raising intriguing moral questions. Is it right to use a deceased person’s image when they can’t consent to it? Isn’t it just plain creepy? Heck, is it even legal?
The lines are pretty thin, and it’s not exactly clear. Another very famous, but not exactly human, performing hologram comes in the form of Hatsune Miku, the turquoise twin-tailed Japanese-based singing program. First debuted by Crypton Future Media back in 2007, she is by far the most famous of the many Vocaloids that have come before and after her — a collective nickname for characters representing synthesized voice software. She’s headlined many sold-out tours around the world — even nagging a spot in the coveted Coachella (though it’s been cancelled due to the COVID-19 pandemic).
Miku and her cohorts differ from resurrecting past singers as holograms in the sense that she’s not real in the slightest, even if her voice samples were provided by a real person (well-known Japanese voice actress Saki Fujita). That doesn’t mean she doesn’t have fans, clearly — Miku strikes a perfect balance of being her own character (of which Crypton created to promote their software) and a collaborative project that’s used by musicians and artists alike. Thousands of songs have been produced using her voice, and she’s been tapped for promoting various products and merchandise.
It’s easy to distinguish Miku from a real person, obviously. She’s clearly not fashioned in the image of someone that we’d see walking down the street. And while the hologram performances of deceased celebrities are appropriately creepy, we know who they still are.
What happens when we can’t tell? Or rather, what happens when the image we’re looking at isn’t real? And it’s trying to sell us something?
Thus, it seems like a natural conclusion — the inevitable intersection of technology, marketing, digitization, social media, and celebrity: virtual, or computer-generated influencers.
Part III: Virtual Beings Overtake Social Media
Arguably one of the most famous virtual stars is Miquela Souza, more commonly known as Lil’ Miquela. She was created in 2016 by a startup called Brud, a company backed by significant money — millions of dollars — in Silicon Valley. She started out simply as a CGI project, but quickly grew a following and soon booked many collaborations and endorsements as though she were real. She also releases music, and has thousands of streams on Spotify. She has a lot of famous friends and celebrities that she takes pictures with. She wants you to know about the Black Lives Matter movement, and that you should vote if you can. She makes YouTube videos where she’s given life thanks to face-mapping and technology. She’s donated hundreds of thousands of dollars to various organizations. And if you happen to scroll upon her work, you’ll most likely oscillate between wondering if she’s real or not.
In truth, she’s an expertly crafted millennial-esque being with a genuine following making genuine money, just like so many others — for the people behind her.
Miquela is among the new wave of computer-generated and virtual influencers that are staking a large claim in not just the advertising market, but the world. So much so that there is even an agency for curating their bevy of virtual models into a portfolio, called Diigitals. Diigitals is home to another rising virtual model Shudu, who booked a gig for Rihanna’s beauty brand Fenty, and subsequently confusing the hell out of people of whether or not she was a supremely beautiful real person or not.
Besides the seemingly everyday CGI model/influencer, the most notable use of this marketing tactic award goes to none other than KFC. KFC is no stranger to weird marketing to get people to eat their chicken — and sometimes, the weirder, the better. Thus, in 2019, they decided to play into the virtual influencer craze and created a computer-generated Colonel Sanders in all of his silver fox glory.
With tongue-in-cheek humor (we all know Colonel Sanders isn’t supposed to look like that, per se), KFC got people talking about them and their brand. They even released a dating sim with an anime-style Colonel to go along with the initiative (you can play it for free on Steam here, if you so wish). Because of its outlandish yet comedic approach, even poking fun at the typical influencer, the Colonel served as a parody that garners headlines and lodged itself in the consumer subconscious.
Part IV: Consequences of the Fake Idol
The virtual influencer is by no means a perfect science — or perfect technology. They’ve evolved as our social media use has, since behind the scenes lies real people pulling the strings behind them to monitor performance and what the latest meme is. They’ve been given larger voices and larger platforms to try and become as humanlike as possible.
Most recently, Riot Games, the creators of MMORPG League of Legends, created lots of hype for their new hero in the form of Seraphine — a pink-haired indie singer that generated a massive following on Instagram and Twitter alike (she’s even verified on both platforms). Seraphine’s response has been both positive and negative, whereas the authenticity is a bit questionable: she’ll tweet about her insecurities and try to show her vulnerable side. She gets nervous before performing. She’s not as confident as she appears to be.
All of it screams, “She’s just like us! She’s just like you!” It’s so lifelike yet unrealistic that it lives well into the uncanny valley. When you look at images of Seraphine, it’s clear she’s not human — but Riot’s marketing team has done their darndest to make it seem like so through her active posting on social media and humane vulnerability.
Of course, there are benefits to utilizing virtual influencers — cheaper and easier to execute, for example. And in the time of a global pandemic, it’s certainly easier to make computer-generated people instead of gathering real ones. But these benefits can be beset by living within a grey area. The most important — and most obvious — notion is that a robot or AI isn’t human, and therefore not held back by the limitations of a typical human. Since their entire personality is manufactured, there’s no worry of mental distress when working with them. Since they can do impossible dance moves humans can’t, or possess impossible forms — there’s been a ploy for alienesque influencers cropping up — there’s no need to worry about overexertion or practice. You won’t have to worry about worker’s comp or safety laws — and you especially won’t have to worry about them getting into scandals (unless, of course, that’s in the plan). A created entity is walking a fine line, dipping its feet into both positives and negatives of not being corporeal. Because despite it not being “real,” it needs to appear as so, or actual humans won’t view them as on par with regular influencers.
You may also walk a line of marketing morality. The Federal Trade Commission has required real-life influencers to hashtag and explicitly state when their posts are paid advertisements, and Instagram has implemented the ability to show if the post is a paid partnership. Even so, there are tons of posts that go unchecked as endorsements and ads, and transparency is harder and harder to find. Are AI influencers held to the same capacity? It’s not supremely clear, though New York Times notes that the FTC “acknowledged in a statement that it ‘hasn’t yet specifically addressed the use of virtual influencers’” and it’s up to the creators to guarantee truth in photo, video, and anything in between.
In the same vein, you have to wonder what the effect of synthetic pixels can do to a generation already feeling inferior to their curated and edited feeds. Miquela’s kind could be perceived as a fascinating commentary on what social media has become today — where it's difficult to distinguish real from fake. With the advent of editing apps and Photoshop, it can be awfully easy for an AI to blend into the crowd. And why not? There are plenty of advantages, and the sheer absurdity of a CGI character doing all the typical tasks of a human — sometimes better than a human can — is surely enough to get eyes on them.
For a young generation often plagued by mental illness, social media helps and hurts. More stories of struggle can be told, and community can be found, yet inundation of perfect images can cause our self-esteem to plummet (despite most being edited). Virtual influencers can complicate this further — despite Seraphine showcasing her skincare routine, she has absolutely poreless perfect skin because she was designed that way. When a human looks in the mirror, they clearly don’t. We can try to distinguish our subconscious from recognizing this, but repeated viewing of images deemed perfect will take its toll, especially on such a vulnerable group.
There is a myriad of ways to make sure that we are consuming social media healthily — limiting our time, reminding ourselves of the innate curation and editing of images, etc. But a feature of the world we live in is being subject to advertising and media, and we must learn to properly distinguish what is suitable. Perhaps a shift on what the norm is considered would prove beneficial. Virtual influencers aren’t going anywhere anytime soon, but neither is our edited world. The more we know something is real — truly real, not edited or manufactured — the better off we may be in our future consumption.
