His winning work “Théâtre D’opéra Spatial” was created with Midjourney, an artificial intelligence system that allows images to be created simply by entering a few pieces of text – for example “an astronaut riding a horse”.
Many artists were angry, but Mr. Allen was unmoved. Mr Allen earned just $300 (£262) from the competition, but the news struck a nerve.
Some artists have already feared that their work could be taken over by a new kind of AI image generator, fearing them to free ride on the years they’ve spent mastering their craft. “This thing wants our job, it’s actively anti-art,” tweeted California film and game concept artist RJ Palmer, which has been liked more than 25,000 times.
In Twitter post, he highlighted how well the AI system’s results can mimic living artists. In one case he investigated, the AI even tried to copy an artist’s signature.
The output of these AI systems is impressive, but they are based on the results of flesh-and-blood creators—their AIs are trained on millions of human-made images.
Stable Diffusion, a recently launched open-source AI image generator, learns from compressed files of “100,000 GB of images” pulled from the Internet, its founder Emad Mostakek said.
Mr. Mostak, a computer scientist with a background in technology and finance, sees Stable Diffusion as a “generative search engine.”
Google Image Search shows you pre-existing images, while Steady Diffusion shows you anything you can think of. “This is Star Trek’s holodeck moment,” he said.
The art of artificial intelligence at a glance
Artists have always learned from and been influenced by others – “great artists steal“, as the saying goes – but Mr Palmer says AI isn’t just about taking inspiration from other artists’ work “That, in a way, directly steals their essence”.
And AI can replicate a style in seconds: “Right now, if an artist wants to copy my style, they can spend a week replicating it,” says Palmer.
“It’s a person who spends a week creating one thing. You can produce hundreds of them a week with this machine.
But Mr Mostaque says he is not concerned about putting artists out of work – the project is a tool like Microsoft’s Excel spreadsheet, which – he notes – “didn’t put the accountants out of work, I still pay my accountants”.
So what message does he have for young artists who are concerned about their future careers, perhaps illustration or design? “My message to them would be, ‘illustration design jobs are very tedious’. It’s not about being artistic, you are a tool”.
He advises them to look for opportunities with new technologies: “This is an industry that will grow significantly. If you want to earn, earn from the industry, it’s more interesting. ”
And there are already artists using AI art for inspiration and money. OpenAI says its DALL-E AI system is used by over 3,000 artists from over 118 countries.
There are even graphic novels made using artificial intelligence. One of the authors called the technology “a collaborator that can excite and surprise you during the creative process.” But despite the outcry over how artificial intelligence uses artists’ work, experts say the legal issues could be complex.
Professor Lionel Bentley, director of the Center for Intellectual Property and Information Law at the University of Cambridge, said that in the UK “usually using somebody else’s style does not infringe copyright”. Professor Bentley told me that artists have to demonstrate that the AI output already reproduces a significant part of their original creative expression in the particular artwork used to train the AI.
Even when it is possible, few artists are able to win such legal battles.
The Design and Artists’ Copyright Society (DACS), which collects fees on behalf of artists for the use of their images, is concerned.
I asked Reema Selhi, policy director of DACS, if the artist’s livelihood was at risk. “Absolutely,” she said
DACS is not opposed to the use of artificial intelligence in art, but Ms. Selhi wants artists whose work is used to monetize AI imaging systems to be fairly compensated and able to control the use of their work. “The artist could not identify the work used in the database and refused,” she added.
When images are pulled from the Internet to be used to train artificial intelligence, artists can claim copyright infringement, although legal experts suggest many reasons why such claims might fail.
Ms Selhi said the proposed changes to UK law would make it easier for AI companies to remove legitimate artists’ work from the internet – something DACS has opposed. Mostak said he understood the artist’s fear and frustration, noting that “you can see it in the photograph as well.”
He said the project is working with “tech industry leaders to create a mechanism for artists to upload their portfolios and request that their styles not be used on online services that use this and similar technologies.”
Deep fakes, pornography and prejudice
Google has previously created artificial intelligence that can create images based on text keywords. Dubbed Imagen “due to the potential risk of abuse”, the trial was never made available to the public. Google warns that image datasets used to train artificial intelligence often contain pornography, reflect social stereotypes and contain “derogatory or otherwise harmful associations with marginalized identity groups”.
The constant spread could be used to create non-consensual pornography, so-called deep fakes, and other questionable images. Mr. Mostak said the unethical use violated the terms of the license. He said the software already filters out attempts to create images that aren’t safe for work, though tech-savvy people can fix the problem. But he said that “someone doing something illegal” was to blame and that other existing tools could also be misused, such as someone using “Photoshop’s blending tool to put a human head on top of a naked body”.
Art or tackiness?
Sci-fi artist Simon Stålenhaag tweeted that the art of artificial intelligence revealed a “kind of derivative, generated goo .. our new tech lords are hoping to feed us”.
There are also some well-known personalities associated with the development of technology. The “King of Technology” Elon Musk is himself a backer of OpenAI.
They say DALL-E is far from “derivative goo” but instead encourages human creativity and creates “unique, original images that have never been seen before.” Policymakers need to set the right rules so that no one feels cheated and that the money isn’t just being taken from artists and going into the pockets of big business.