But the law works on a spectrum and any involvement of machines does not invalidate the ultimate authorship of a human being. “Obviously, when I use Microsoft Word to write a book, it’s just a tool and the expression is defined by me,” Weatherall says.
She is convinced that most AI-generated images are on the wrong side of the copyright rule: too many machines, too little human creativity. “If you write ‘I want a picture of Boris Johnson with fish coming out of his ears’ and it generates a picture that looks like this, then that’s interesting in that spectrum I described earlier. It’s the human who makes choices – I want Boris Johnson, I want fish, I want it in his ears – but the expression is really generated by the system.”
One of its hallmarks is style. DALL E Mini, which has swept the internet in recent weeks, produces grids of images that are instantly recognizable, with fuzzy, visibly digital renderings and distorted faces. Users can also specify a style such as “digital art” or “woodcut” in DALL E 2 and Google’s imagebut they are limited by the underlying data set and a user’s ability to reduce an aesthetic to a brief description.
But even if purely AI-generated images are not art as recognized by copyright law, they will affect the art world. Many artists have created their own AI tools to visualize data, or have transformed images originally created by AI into their own work. Creating a quick logo for a company, a digital illustration for an opinion piece, or framing a cartoon probably seems to become trivially easy as well. That doesn’t bode well for people doing graphic design and animation work, who could be pushed further up the value chain to correct work initially done by AI.
Another consequence could be a flattening of the style. The internet is already full of futuristic, laser-eyed and steampunk-esque imagery that has been particularly associated with non-fungible tokens, an online ownership tracking system that should cover any genre of digital images. AI images could anchor that.
But Ellen Broad, an associate professor at the 3A Institute in the Australian National University’s cybernetics school, doesn’t believe in the most apocalyptic statements. “Do I think this is the end of human creativity and expression? No.”
“In three years’ time, when everyone is using the same kind of image generation models, there’s going to be a market… for something that looks different,” she says.
Broad may well be right. But AI has a long history of fooling people into seeing deeper meaning in the output. Blake Lemoine, the Google engineer, was captivated by the poetic but nonsensical responses his company’s chatbot LaMDA generated when he asked about its soul. “I see my soul as something akin to a stargate,” LaMDA said, according to a transcript Lemoine published online after his resignation. Funerals are held for dismantled dog robots released by Sony in the 1990s.
“It’s very easy to anthropomorphize,” said Jasmin Craufurd-Hill, an emerging technology researcher and director of advanced technology at the Australian Risk Policy Institute. “People have connected and started to assign human characteristics and behavior to our technology.”
Still, Imagen and DALL·E 2 don’t show realistic people for the time being. “There’s a reason there aren’t people,” says Craufurd-Hill. “And it relates to these incredibly problematic data sets.”
Many large datasets, on which AI systems often draw, contain images that are racist, sexist or inappropriate, such as pornography, says Craufurd-Hill. If an AI is trained in such a database without the proper guardrails, they can end up returning the same kind of problematic material, even if users don’t use it maliciously.
In an elliptical confessional essay 2015 when explaining his Instagram-based exhibit, Prince seemed to predict the troubling no-man’s-land in which AI has arrived.
“The ingredients, the recipe, ‘the manufacture’, whatever you want to call it … was known, but had turned into something I had never seen before,” he wrote of his works. “I wasn’t sure if it looked like art. And that was the best part. Doesn’t look like art. The new portraits were in that gray area. undefined. In between. They had no history, no past, no name. A life of its own. They will learn. They will find their own way. I have no responsibility. They do. Friendly monsters.”
A cultural guide to going out and loving your city. Sign up here for our Culture Fix newsletter†
#Photorealistic #graphics #arrived #artists #trouble