Artists’ lawsuit against Stability AI and Midjourney gains momentum

Many artists have claimed that popular generative AI services have violated copyright law by training on a dataset that included their work, and in some cases, users of the services can directly reproduce copies of the work. Last year Judge William Orrick allowed a direct copyright complaint against Stability, the operator of the popular Stable Diffusion AI image generator. But he dismissed a number of other claims and asked the artists’ lawyers to supplement them with more details.

In this newer ruling, the revised arguments convinced the judge to approve an additional claim for aggravated copyright infringement against Stability. He allowed a copyright claim against DeviantArt, which used a model based on Stable Diffusion, and against Runway AI, the initial startup behind Stable Diffusion. He also allowed copyright and trademark infringement claims against Midjourney.

The latest claims include allegations that Midjourney misled users with its “Midjourney Style List,” which listed 4,700 artists whose names could be used to generate work in their style. The artists argue that the list — created without their knowledge or consent — constitutes a false endorsement, and a judge found the accusation significant enough to warrant further argument.

Judge Orrick remained unconvinced by some of the arguments, which he had previously referred back for more detail. He dismissed claims that the generators violated the Digital Millennium Copyright Act by removing or changing copyright management information. He also dismissed a claim that DeviantArt violated its terms of service by allowing the scraping of user works to obtain AI training datasets. And of course, the claims he allowed will still have to be presented in court.

Kelly McKernan, one of the artists who created the costume, described the ruling as “very exciting” and a “HUGE win” for X. McKernan noted that passing this initial step allows them to demand information from companies during disclosure — potentially revealing details about software tools that often remain black boxes. “Now we can find out all the things these companies don’t want us to know,” McKernan wrote. (If companies If (If the information were ordered to be provided, it would not necessarily be released to the public.)

But the outcome of the case is tough to predict. Numerous lawsuits have been filed against AI companies, claiming that tools like Stable Diffusion and ChatGPT easily reproduce copyrighted works and are illegal trained on massive volumes of them. The companies have responded that these reproductions are sporadic and arduous to produce, and argue that training should be considered legal fair apply. Some early lawsuits have been dismissed, including the GitHub Copilot case, whose dismissal was cited in yesterday’s ruling. Others, such as the Fresh York Times Company’s lawsuit against OpenAI, remain pending.

At the same time, OpenAI, Google and other tech giants have struck deals with publishers (including Edge parent Vox Media) and image providers to ensure continued access to the data. Tiny companies like Stability and Midjourney have less money to buy access to the data, and individual artists have less ability to demand payment — so the legal stakes are especially high for both sides of this dispute.