In December of last year, variations on a simple image — a general prohibition symbol crossing through the word AI with “NO TO AI GENERATED IMAGES” beneath it — began popping up on my Instagram feed. I had just written my last piece for Hyperallergic, looking at Artificial Intelligence’s capacity to relay and become part of art history, and I realized that the interviews I had conducted for the article had all been with pro-AI voices. I had not spoken to the group being hurt at the hands of AI image generation: digital artists. The recent popularization of AI software such as OpenAI’s DALL-E 2 and ChatGPT has made AI the topic of the moment. AI-generated images, from quirky profile pictures showing users as fantasy characters, to reimaginings of famous works from art history, were popping up everywhere, and questions about the technology’s future use and current methods quickly grew. When an advert for AI copywriting service “Jasper” came up on my feed, I, too, became concerned about the future security of my work as a writer in an ever-technologically advancing landscape.
AI art is generated using software now readily available to online audiences such as DeepAI, DeepDream, and OpenAI’s DALL-E 2. These programs have been trained to create images from text prompts, using millions of pairs of images and captions from the internet to “‘learn” from, and becoming more accurate at the process with every interaction. The images used in the datasets are publicly available on the internet, but Getty Images has started its legal proceedings against Stability AI for breach of IP rights, and other lawsuits are brewing.
The key issue people seem to be taking with AI art is that the artists who created the images from which the programs were trained were not consulted and are not remunerated for their work. Digital artist @loisvb wrote in a post on December 15: “I get zero compensation for the use of my art, even though these image generators cost money to use, and are a commercial product … platforms should do everything they can to prevent scraping of their content for these databases.”
News broadcaster and arts consultant Robyne Robinson explained this issue of “scraping” to Hyperallergic, saying: “’Scraped’ and reconstituted art is basically theft of an artists’ work. It undermines the hours, weeks, and months artists put into conceiving and creating their work and has the ability to disenfranchise them in a matter of minutes. AI can undercut a skilled artist’s economic viability.”
Dutch artist @the.art.fish in a December 19th post compared the issue to illegally downloading music (a landmark court case which took place in 2010, finding the creator of LimeWire — a free music download software believed to be installed in over a third of all computers around the world in 2007 — guilty of copyright infringement and inciting others to do the same), asking “why is our work as visual artists any different?”
On December 5, 2022, the first ‘No To AI-Generated Images’ post was made by the Bulgarian illustrator Alexander Nanitchkov, accompanied by the #notoaiart hashtag. Nanitchkov’s experience with the issue began on ArtStation, a platform for digital artists which was founded in 2014 and has over 3.4 million users worldwide. ArtStation was accused of not doing enough to prevent artists’ work from being scraped from the site, and for not suitably flagging when work uploaded to the site was AI-generated. Nanitchkov spoke out for artists whose labor had been used free of charge in the creation of AI art, which he wrote in the post is “created on the backs of hundreds of thousands of artists and photographers who made billions of images and spend time, love and dedication to have their work soullessly stolen and used by selfish people for profit without the slightest concept of ethics.”
Nanitchkov also warned about AI datasets at large, writing, “Do not think that this concerns only artists, it is a breach of private data, medical records, pictures of your kids and family, pornographic content … it has Everything.” The artist has now altered his ArtStation profile so that all images are watermarked, many reading “create don’t scrape.”
AI art appears to be a particular threat to digital artists. Part of this is because these creators tend to produce large bodies of content, meaning that datasets will have access to huge amounts of information on their style. The art historian and curator Alexandra Steinacker gave me examples of fine artists who have been excelling at incorporating AI into their practices, asserting that AI “has been playing a role in the art world that not only refrains from taking away from a human artist, but actively adds to their practice.” This may be true for artists working in physical media who can transform AI-generated images in a non-digital space, but the boundary between digital art and AI-generated art is much harder to draw.
While AI art may be a more obvious threat to digital artists, others argue that it is a threat to creatives across the board.
“Independent working artists already have enough hardship to deal with,” wrote American artist Matt Verges when sharing Nanitchkov’s original image. “AI in the workforce is not an issue only relegated to the art world … And you better believe that CEOs are already planning ways to cut their bottom line using this technology, and further exploit their workers.”
Many fine artists work in corporate design and illustration to financially support their personal practice, so the impact of AI taking over UI and graphic design roles could be ruinous for them. This would only serve to make creative careers inaccessible to those without independent financial security; only 7.9% of actors, musicians, and writers are from working-class backgrounds in the UK.
AI is here to stay, and is spreading into more and more areas of our lives. The answer to — or at least some pain relief for — the issues around AI art appears to be new legislation. Senior machine learning engineer at Twitter Rishabh Misra told Hyperallergic that “such regulations could take many forms including watermarks in the images generated by AI, making AI art generation tools paid, developing government policies around enforcing fair use of such content” in order to “curb the societal harm AI-generated art may cause.”
There are some who welcome the dawn of AI art. English new media artist Harry Yeff said that while this technology will make “many writers, artists, composers, and designers simply obsolete, many of these tools are technically better than 60% of ‘creatives’ in these industries.” This, he continues, will “force artists to reinvent what art is. You can embrace and take ownership of its potential, or simply be left behind.”
Art historian Ella Nixon told Hyperallergic that she was hopeful that the boom of AI-generated art will “provoke a new Romantic movement.”
“Artists will revolt and assert their creativity as an inherent human capacity,” Nixon said. “In this regard, I think the physical art world is about to become a lot more interesting.”