Type in a prompt like “a chocolate bar riding a bicycle in the style of Picasso,” and artificial intelligence tools including DALL-E, Midjourney and Stable Diffusion can conjure an image for you in seconds. They do so by incorporating elements from the vast libraries of digitally available images and artwork from across the internet that they have been trained on.
But in doing so, are those A.I. tools infringing on the copyrights of the artists behind those images? That question is at the heart of two new lawsuits.
Last week, Seattle-based stock image giant Getty Images announced that it has initiated legal proceedings against Stability AI, the maker of Stable Diffusion. Getty alleges that the company has copied millions of its images and “[chosen] to ignore viable licensing options and long-standing legal protections in pursuit of their stand-alone commercial interests.”
Stability AI was also named as a defendant in a class action lawsuit filed in a United States federal court in San Francisco by three visual artists on behalf of the visual arts community, per Reuters’ Blake Brittain. They say that A.I. image generators “violate the rights of millions of artists,” also naming Midjourney and DeviantArt, an online gallery that has launched its own A.I. tool, as defendants.
Both lawsuits argue that by scraping the web for images, A.I. image generators unjustly rob artists, using their work without crediting or rewarding them.
“The harm to artists is not hypothetical—works generated by A.I. Image Products ‘in the style’ of a particular artist are already sold on the internet, siphoning commissions from the artists themselves,” the class action lawsuit states. The plaintiffs “seek to end this blatant and enormous infringement of their rights before their professions are eliminated by a computer program powered entirely by their hard work.”
Most image generating services charge a fee, including Midjourney, DALL-E and Stable Diffusion.
In response to the artists’ lawsuit, a spokesperson for Stability AI says the company takes “these matters seriously,” and that “anyone that believes that this isn’t fair use does not understand the technology and misunderstands the law,” per Reuters. Speaking with the Associated Press in December, before the lawsuits were filed, Midjourney CEO David Holz compared the process behind his image generating service to the process behind human creativity, which often entails drawing inspiration from other artists.
“Can a person look at somebody else’s picture and learn from it and make a similar picture?” Holz said. “Obviously, it’s allowed for people and if it wasn’t, then it would destroy the whole professional art industry, probably the nonprofessional industry too. To the extent that A.I.s are learning like people, it’s sort of the same thing and if the images come out differently then it seems like it’s fine.”
But the plaintiffs are not alone in their concerns. In December, an A.I. photo app called Lensa went viral for its new Magic Avatars feature, which took selfies uploaded by users and transformed them into an assortment of artistic renderings. The feature prompted concern and criticism from the arts community, especially because some Lensa images displayed mangled versions of what appeared to have once been artists’ signatures.
Matthew Butterick, a lawyer working on behalf of the plaintiffs in the class action lawsuit, filed another lawsuit last year against GitHub Copilot for its “unprecedented open-source software piracy.”
“Since then,” he writes in a blog post, “we’ve heard from people all over the world—especially writers, artists, programmers and other creators—who are concerned about A.I. systems being trained on vast amounts of copyrighted work with no consent, no credit and no compensation.”