When Jason Allen won the digital-art competition at the Colorado State Fair last year, he sprayed fuel on a debate about the role of artificial intelligence in the art world.
Now the Pueblo-based game designer, who created his award-winning piece “Théâtre D’opéra Spatial” using the AI software Midjourney, is exploiting his fame as an AI-art poster child to launch a campaign to legally protect AI works.
“The U.S. Copyright Office rejected my copyright registration (for the image), so after some back and forth, I’ve hired a lawyer and am appealing,” said the 39-year-old Allen, who this week is unveiling a coordinated online protest against the ruling. “We’re prepared to go all the way to the Supreme Court.”
Allen’s Colorado-based protest is called COVER, or Copyright Obstruction Violates Expressive Rights. He’s filing a Request for Reconsideration with the U.S. Copyright Office in an attempt to establish sole ownership of an artwork generated using AI software — the first appeal of its type, he said. It parallels international debates and legal cases about revenue and commercial rights with AI creations, but takes specific issue with the U.S. Copyright Office’s reasoning.
“We have decided that we cannot register this copyright claim because the deposit does not contain any human authorship,” Copyright Office officials wrote in their decision. “Instead, the deposit contains only material that your client solicited from an artificial intelligence art-generator.”
In response, Allen and Denver-based trademark attorney Tamara Pester argue that “the use of AI in the creation of art is a legitimate form of artistic expression,” and that such works should be afforded the same protection as “traditional” forms of art.
Allen’s public relations campaign, and his plan to repeatedly appeal any rejections, echoes others in the art and design worlds who have come to see AI as a tool rather than a threat. That’s in contrast to the backlash Allen received from artists around the globe after “Théâtre D’opéra Spatial” won in Colorado last year.
Many said that his piece stole recognizable imagery from existing paintings and photographs, and that his creative process was a cheat, relegated only to a few keyboard clicks. But others disagree, or at least see AI as a groundbreaking technology that could challenge our assumptions of what art can be.
“I believe that as artists, we have the power to imbue these creations with meaning and emotion,” said Denver artist Mario Zoots, who sees endless possibilities in AI for his collage work. “As Baudrillard wrote, ‘Simulation is not a copy of the real, but the hyperreal.’”
AI imaging software uses text-based prompts to scan and synthesize information from large databases of imagery, such as Getty Images, which licenses stock images and photographs. Getty filed a lawsuit last month in London against the AI program Stable Diffusion for “brazen infringement” in what it describes as the theft of 12 million images from its collection.
Artists Kelly McKernan, Sarah Andersen and Karla Ortiz in January filed a class-action lawsuit in San Francisco against the AI program that Allen used — Midjourney — and its peers, Stable Diffusion and DreamUp, for stealing their original, copyrighted work for “collage” purposes in AI engines. They found that users were specifically including their names (and therefore sampling their work) in the search terms and generation of AI imagery, which could then be shared or sold without credit and revenue, The New Yorker reported in February.
“(Artists) are acting like it’s this massive, end-of-the-world event,” Allen said. “But the backlash and animosity is just an outlet for them to express their frustrations about everything else, since it’s not polite to criticize other artists.”
Even as many social media debates and op-eds continue to oppose the idea of copyrighting AI-generated art, Allen remains the willing lightning-rod, doing nearly 100 interviews since his state fair win and declaring to The New York Times: “Art is dead, dude. It’s over. AI won. Humans lost.”
But while Allen is litigating his case with the public, AI is continuing to evolve in unexpected ways.
“There are so many unanswered questions regarding the ethics, legality, and societal impact that will undoubtedly be the focus of conversation in the coming months,” said Cody Borst, who has worked on tech-forward installations as director of exhibition at Meow Wolf Denver, an immersive art company. “The speed in which the technology is advancing is quickly outpacing our understanding of the impacts it might have.”
Machine learning, or data training, as some AI processes are also called, has gone mainstream. Dalle-E 2 users have shared absurd imagery from their text-to-image experiments, while the Lensa AI app has led a wave of smartphone users to post uncanny AI portraits of themselves on social media. ChatGPT continues to generate alarmingly sophisticated yet soulless, thoughtless text ranging from business plans and homework to beer recipes and poetry.
Fundamentally, there’s no hard line between handmade and technologically aided art, said Sharifa Moore, executive director of the Denver Digerati nonprofit, which in September will hold its retooled Digerati Emergent Media Festival (formerly Supernova).
“There are so many precedents that have blurred that line throughout art history … Modern art introduced the idea of art being as, or more, important than the tangible thing,” she said. “Even photography had to fight for respect at the dawn of the 20th century.”
AI also has limitless applications for filling in blanks in the world of imagery, said David Romero, a community manager at Portland, Ore., company Luma Labs. His company is looking to “democratize 3D,” according to its website, and said AI has already helped sharpen scans of Colorado landmarks such as East High School and Empower Field for use in 3D digital environments.
“Each one of these scans would have been 100 percent impossible to create a year ago,” said Romero, a graduate of East High. “To make something remotely similar with previous technologies would have cost (more money) per scan and only have a fraction of the photorealism. Now it takes less than 10 minutes and is basically just the cost of cloud compute time.”
Colorado’s Allen will soon unveil an ambitious project that delves into the creation of his award-winning Colorado State Fair piece, he said. If people understand how current AI works, its perceived threat would fade, he said.
“It literally at this point is a calculator,” Allen said of AI image-generating software. “It’s a very complex matrix calculator running through a model that was trained on all these images. But if you were a person of capable of looking at all these images, you could do it too.”
