When award-winning artwork can be generated by algorithms, where does that leave artists?
By Tiana Cline
What is an artist? In the age of artificial intelligence, machine learning and text-to-image tools, it’s questionable. Anyone with a computer can generate stunning images in seconds thanks to tools like Midjourney, DALL-E, Stable Diffusion and Craiyon. These algorithmic programs, which work with text prompts and data sets, are already starting to shift the creative economy. In August 2022, an American game designer named Jason Allen won first place at the Colorado State Fair Fine Arts Competition. His award-winning work – which has been described as ‘Renaissance meets steampunk’ – was created using Midjourney. And even though the category in which Allen came first was ‘digital arts/digitally-manipulated photography’, it left a lot of people in the creative world angry, questioning whether this kind of AI-generated art could replace human artists.
“There’s no doubt that the whole visual world is totally oversaturated with imagery. I know a lot of people who do AI art. I see it on Instagram or Facebook and some of the images have stuck in my head,” says Roger Ballen, an American artist who lives in Johannesburg, South Africa. “So, I think they’re successful from the point of view of conveying an aesthetic message from one person to another. And how they were able to achieve this, is this important or not important?”
For some industries, AI-generated art is simply a natural progression of where we’re heading – it’s a key part of Web3, the metaverse and how AI algorithms can be integrated into different industries. According to Midjourney founder David Holz, business owners, game designers and people in the movie industry use the tool for storyboarding, strategy and content development. But that’s not where AI-art ends. The Economist used Midjourney to produce one of its 2022 covers and The Atlantic chose AI-generated art for an editorial in lieu of hiring an illustrator – going so far as to credit Midjourney as the artist.
For Ballen, AI-generated art is similar to a computer beating a world chess champion – it’s inevitable. “You look at the work that’s been done and some of it is amazing. This is where we’re going – going into a world where our lives and viewpoints and experiences are more virtual rather than physical and it’s happening at an accelerating pace, he says. “I’ve been doing photography for 54 years now. When I started, you needed to go to university and there was a lot of training that took place to get there. Now anyone can do photography. For one photograph that was produced 30 years ago, there may be 50,000 now.”
Plagiarism by machine
There are many ethical considerations around creating (and selling) AI-generated art. AI often reflects human bias which means that the images contained in online datasets are subjective to what is available online.
Ballen questions being able to put a price on an artwork when the material comes from endless sources. He also asks if there is a difference between using an image of a tree (as one example) or from a well-known artist. “Some of these people are trying to produce works of art but actually, the work itself isn’t necessarily theirs – it’s an amalgamation of other people’s work compiled by a computer. So, the issue then is can that person sell it as their own when it’s ultimately other people’s work?”
Joh Del, a Cape Town-based artist and illustrator’s understanding that that all artists learn from other pieces of art – both human and AI. “Every time you put something out into the world, someone else can see it, and learn from it. I think almost all species of animals on earth learn from each other,” says Del.
Mikhael Subotzky, an award-winning artist based in South Africa, doesn’t see AI-generated art as something fundamentally new or different. “I think that the power and relevance of art is that its boundaries are always expanding – there should be space for every kind of art,” he says. For Subotzky, the trickiest issue is copyright. “There really should be a way for artists to opt in or out of the algorithms learning from their work. Or even a royalty system where they get a cent or whatever for every time their work is used in this way. Of course, this would require a fundamental change in how images circulate. I wish that the blockchain crowd would focus on this type of thing rather than the latest NFT prices.”
For now, the law is yet to catch up with the advances in AI technology. Under copyright law, copyright depends on human input and AI-generated art is made by machine. In order to avoid future legal issues, Getty Images has banned the sale of AI generative artwork created using image synthesis models. “In one way, it’s an image that somebody initiated to put on a wall and show other people and on the other hand, it’s a mixture of other people’s property. In the commercial world you have copyrights, you have patents, you have rights of property and with this, it’s not as clear because it’s a mishmash of thousands of things that are put together into an artwork,” says Ballen. “You have to look at the image itself. You have to deal with the aesthetics in the image, the message in the image, the type of artwork, where the information came from, how it was put together – there’s all sorts of things that make it almost impossible to come up with a clear answer… where this ends and where it begins, I don’t know.”
But Is It Art?
“I’m the type of person who believes in the autonomy of the work, that you don’t necessarily need to know anything about the artist, the place, the time – anything,” says Roger Ballen. “It doesn’t matter whether you know anything about anything. Art should be like coming into contact with a flower. All of a sudden, you smell it, you feel it, the color leaves an imprint on you and you don’t necessarily need to be a botanist, all that matters is the experience of the art itself.”