The TikTok challenge How AI art won social media

Cover Image - The TikTok challenge
Óscar Raña
WordsNeelam Tailor

In some ways, AI is to the 21st century what the camera was to the 20th: a shiny new tool that enables hitherto impossible creations. And like photography, AI-generated art is a major draw on TikTok and other social platforms, where followers are flocking to see novel prompts brought to life. But with issues of artistic consent and misinformation looming, there are important risks to be considered, too. Neelam Tailor speaks to some of the artists at the forefront of TikTok’s AI art scene to discuss the controversies and possibilities the new medium presents.

Artwork by Óscar Raña.

Italian surveyor Fabio Comparelli has always dreamed of making a living from his artistic creations. “About seven months ago, I discovered AI art, and it felt like an explosion in my mind. It allowed me to express ideas that I had for a long time,” he says. 

Under the username, the 34-year-old posts arresting time-lapse animations to an audience of over 300,000 on Instagram, and 31,000 on TikTok. His most famous video depicts past, present and future stages of human evolution, with a primate progressing through the human phases, before assuming increasingly robotic forms. Some commenters bristled at the concept of an algorithm generating a picture of humanity’s future, but as Comparelli assured his followers, it was he who chose the prompt that created every frame of the video.

Since then, Comparelli has created more visually and emotionally impactful animations. In “Deconstruction,” his thought-provoking video on human-induced climate change, a family of orangutans sit in a tree that becomes a burning forest, which is then destroyed by humans to make way for concrete cities. “I enjoy the challenge of creating coherent animations and telling a story with AI art,” he says, noting that it takes “a lot of learning, trial and error, questioning, and perseverance” to create something meaningful. In the future he wants to focus on addressing social, political, and environment issues in this way.

Since generative AI image-making tools went mainstream in 2022, social media has been flooded with such visualizations and convincing deepfakes, posted by a new breed of tech-savvy creators. “When I was experimenting with the old technologies—which, bear in mind, was only a few months back—I was not very convinced [about AI art]. But more recently, I got access to DALL-E and Midjourney, and that’s another level,” says 21-year-old British artist Lola Holliday.

Holliday, who has a background in traditional art, is one of a growing number of AI-generated content creators on TikTok, where she’s become popular for visualizing phobias. The platform has been instrumental to the spread of AI art; and through interacting with creators like Holliday and inputting their own prompts in the comments, followers almost create the art themselves.

“Even though obviously it's not a real person with a real brain, there is some aspect of it that feels like you're really picking at the brain of something to see what it interprets,” says AI artist James Musto. Posting as Gaxalactic on TikTok, where he has nearly 73,000, Musto made a name for himself with AI-generated images of gods of different countries.

Join the club

Like this story? You’ll (probably) love our monthly newsletter.

While decades of film and television have prepared us for a future where machines take over menial and mundane tasks, I don’t remember seeing the one where machines started writing viral hits and producing otherworldly imagery. 

Journalist Chris Stokel-Walker, the author of “TikTok Boom: The Inside Story of the World's Favorite App,” however, sees the common conception of artificial intelligence as an “incorrect use of terms.” The world has a tendency to anthropomorphize this technology, he explains, which doesn’t help with understanding the realities of it. “We use the term ‘generation,’ which kind of makes people think of creation, where in actual fact it's recreation or collaging,” he says. “Essentially, AI is a really fancy, very technical, well-disguised collage. It is taking past works, and playing with what they look like, and re-presenting that content to us.”

The data sets used to train AI image-making tools have been a particular point of contention. Some platforms obtain permission from every artist included in their sets, but the most popular programs, Midjourney and DALL-E, were both trained on LAION-5B—a dataset of 5 billion images downloaded from the internet without their creators’ consent. “Traditional artists’ work has ended up in these data sets,” says Isabelle Doran, CEO of the Association of Photographers in the UK. “Has the machine already got to the point where it's learned to mimic [an artist’s images], even if they request those images be removed?” I asked artificial intelligence to turn phobias into art… do these resonate with you? Which phobias should I use as prompts next? #ai #aiart #midjourney #midjourneyart #art #digitalart #phobia #arachnaphobia #acrophobia #mentalhealth #aigeneratedart ♬ original sound - 🦅

Doran describes the creative industries as the “canary in a coalmine” of the generative AI revolution, and urges governments to protect traditional artists. But as Musto says, “You can’t put the genie back in the bottle.” Legislators are still struggling to contain the sprawling World Wide Web 30 years since it was made public, and generative AI could cause an even more formidable challenge. As courtrooms and congresses scramble to answer the ethical and philosophical questions around intellectual property ownership, the protection of people’s livelihoods, and the greater public good, the spread of AI images, music, and text continues.

This explosion has also caused many to revisit the age-old question: What is art? Many traditional artists have dismissed the concept of AI art by definition, which graphic designer and AI artist Edward Sun believes parallels the early skepticism around photography. “I view AI as a tool not so different from cameras,” he says. “As long as the person using the tool wants to express something, I think it is art.”

Dead End Gallery in Amsterdam, which bills itself as the world’s first physical AI art gallery, takes this a step further. The team uses AI to create different artists, and then programs them each to create artworks to display in the gallery space. “The definition that art must be created by humans is long gone,” say founders Constant Brinkman and Paul Bookelman. “Besides, we, the gallery owners, are talking with our AI artists on a daily basis, therefore there is human interaction.”

Tweet this
“I view AI as a tool not so different from cameras. As long as the person using the tool wants to express something, I think it is art.”
Edward Sun

The pair says art is defined by the viewer, something that Comparelli agrees with. “I would argue that the emotional impact of art is not tied to the medium or process used to create it. What makes art genuine is the emotional connection the viewer feels with the piece”. 

In April 2023, the Sony World Photography Awards unknowingly awarded a prize to an AI-generated image submitted by German artist Boris Eldagsen, who refers to this new category of art “promtology”—“It is related to the practice of creating works by prompts, and encompasses images that look like photography, drawing, painting, objects, moving images, and also music,” he says.

In a way, AI art is a great equalizer, giving people who can’t draw or produce music the chance to make their ideas a reality. “There are a lot of things that I can't create without having a skill of fine art,” Sun says. “AI kind of closes that gap. It makes it quicker for you to go from idea to execution.” 

Does it therefore devalue the efforts of people who have built up those skills over decades? Eldagsen, who is 52 and has been an artist for 30 years, says absolutely not. “For the first time in a technical revolution, older people have an advantage because they can use all the knowledge and skills they collected. I would say to all the artists who are afraid, use your knowledge and combine it with AI. It’s going to bring you to another level,” he says.

Stokel-Walker believes the fascination with generative AI is in part the result of decades-long hype—since the 1950s, AI has been sold as a coming attraction, just five to 10 years away. “We've been sold that false promise, and it’s finally here,” he says. 

He also believes TikTok has a big role to play in its popularity. “In the same way that TikTok revolutionized the act of [video] creation, and made it much more democratic because you didn't need big lights and cameras, I think it’s also helping make people aware of how easy it is to create stuff beyond video through generative AI.” 

To Holliday, generative AI is a convenient way to satisfy TikTok’s relentless demand for new content. “As soon as something is posted, after a day or two, it's now old. If you want to build any kind of social media presence, you have to be consistent and keep putting out really good quality stuff. And I think that because AI art can be pumped out very quickly, it’s got an advantage in that scene,” she says.

Everyone seems to agree that AI-generated images come with risks. “I think the more sophisticated the technology becomes, the more damaging it will be for humanity, on a micro scale—as an individual, comparing yourself with AI-generated people may impact your mental health. But then, on a broad scale, like we saw in America's elections, because of the control over information that social media has.” says Sun.

“You’ve got to be really careful with anything you see online today, because it could be generated by AI,” says Musto, who admits that he’s often duped by photorealistic fakes himself.

Eldagsen entered his AI photo to the photography competition to spark discussion around such issues. “I wanted to test the system to see if it’s prepared for AI-generated images. I found out it is not,” he says. “We need to come up with a system for the news media to [confirm whether] an image is authentic, manipulated, or [AI-]generated, so we can ensure that there is a basis of fact… The problem is that the news media need time, money, and staff to do all that fact-checking, but they don’t have it.”

Beyond misinformation, Sun believes the issue of compensation for artists will need to be addressed, too. “There needs to be a payment structure that becomes standardized for people whose work is being used to train these models so that it can also benefit artists,” Sun adds. 

The next generation of artists is also a concern. “Someone who's born 10 years from now, would they ever bother taking up art as a hobby if the computer can just do it for you?” asks Holliday. 

When asked how humanity should respond to this genie freed from the bottle, Stokel-Walker says: “You double down on the human. You think about what it is that is unique about yourself and the essence of the human experience.”