Getty Images, the world’s leading visual communications company, is suing Stable Creations, the creator of an AI image generator, over allegations of data scraping.
The lawsuit was recently filed in California, and it claims that Stable Creations, which was created by Ahmed Elgammal and Kenneth Boucher, violated the Digital Millennium Copyright Act by “unlawfully accessing and copying” Getty images without permission.
According to the complaint, Stable Creations’ AI generator, known as “Stable Diffusion,” was designed to automatically identify and extract images from Getty’s library and reproduce them in its own original works of art. The generator first searches the web, including Getty Images’ library, for images that match the requested image and then manipulates them in a way that it deems “artistic” and “novel,” ultimately creating a new and unique piece of digital artwork.
The lawsuit alleges that Stable Creations’ use of Getty Images’ library for its AI generator constitutes a form of “data scraping,” which is when a website or program automatically copies and stores information from another website without permission. Getty Images’ Terms of Service explicitly prohibit data scraping and similar activities, meaning that Stable Creations’ use of Getty Images’ library was a violation of copyright law.
In addition to the data scraping allegations, Getty Images is also claiming that Stable Diffusion “infringes upon the exclusive rights of copyright owners,” meaning that by reproducing images within its own artwork, Stable Creations is violating the rights of the original copyright holders.
Stable Creations argued that its AI generator is “transformative” and should be protected under fair use law. However, the court rejected this argument and ruled that Stable Creations’ “re-purposing” of images “without any authorization or license from the copyright holder” constitutes copyright infringement.
Although the lawsuit is still ongoing, it serves as a warning to all creators of digital artwork and AI tools that if they are making use of existing copyrighted imagery, they must be sure to obtain the proper permissions first or risk facing potentially costly legal action.