- A recent court ruling has said that the copyright infringement lawsuit filed by artists against a few GenAI companies can proceed toward discovery.
- Artists have celebrated the ruling, claiming that it would allow them to gain more information about the workings of these companies’ software tools, which are usually black boxes.
In September last year, a few artists sued OpenAI and other generative AI (GenAI) developers, such as Stability AI, Meta, Microsoft, and Midjourney, for illegally using their work to train AI models. According to the latest court ruling, the lawsuit can proceed further toward discovery, with a few claims being dismissed.
A Brief Background
Since GenAI models’ and tools’ capabilities started growing and they could produce content beyond text, a battle has ensued between artists and GenAI development companies. Many artists have accused companies of using and scraping their works to train AI models without permission. Several GenAI companies have come under fire for alleged copyright infringement. This has even led other organizations and institutions, such as the University of Chicago, to develop tools like Nightshade and Glaze to prevent AI models from scraping digital artwork data and make it challenging to reproduce original works.
Judge Allows Amended Claims To Proceed
Last year, a few authors and artists filed a lawsuit against GenAI developers, such as Stability AI, OpenAI, and Midjourney, for copyright infringement. They alleged that these companies used their works without authorization and trained AI systems to respond to human prompts. They also said that users of these services could sometimes directly reproduce copies of their work.
Following this suit, Judge William Orrick allowed a copyright infringement complaint against Stability AI while dismissing a few other claims. He also asked the plaintiffs’ attorneys to amend the claims with more detail.
In a recent hearing, the judge reviewed the revised arguments and approved an additional claim of induced copyright infringement against Stability AI. He also allowed copyright claims against DeviantArt and Runway AI and copyright and trademark infringement claims against Midjourney.
The claims against Midjourney specifically alleged that it misled users with a “Midjourney Style List” that included 4,700 artists whose names could be used to generate works in their style. The artists argued that the list was created without their knowledge and implied a false endorsement.
The judge found the allegations substantiate enough to merit further argument. That said, he quashed some of the arguments he had earlier sent back for amendments. For example, he dismissed allegations that the companies violated the Digital Millennium Copyright Act of 1998 and a claim that DeviantArt had breached its terms of service by allowing users’ work to be scraped to train datasets.
Artists Celebrate the Ruling
While the ruling is just an initial victory, and the case still needs to be argued in court, many artists expressed excitement and welcomed it. Kelly McKernan, an artist behind the lawsuit, said that the ruling would allow them to request more information from companies in discovery, which could unravel details about the software tools that are usually a black box.
Another artist, Karla Ortiz, said in a post on X that the ruling also meant companies using Stable Diffusion models for and/or large-scale artificial intelligence open network-like (LAION) datasets may now be liable for copyright infringement violations.
A High Stakes Case for AI Companies and Artists
The case’s outcome is difficult to predict. Many lawsuits have been filed against AI companies earlier, alleging them of easily reproducing copyrighted works and illegally training their tools like ChatGPT and Stable Diffusion on huge volumes of such content. However, AI companies have been successful in defending themselves many times. Several early lawsuits have also been dismissed, while some, like the New York Times suit against OpenAI, are ongoing.
On the other hand, several big companies like Microsoft and OpenAI are striking multimillion-dollar deals with publishers and artists for ongoing data access. However, smaller companies like Midjourney and StabilityAI have less money to buy access to data, and artists have less leverage to demand payments.
Further, the widespread adoption of AI in moviemaking or other such creative processes will depend mainly on how courts rule in such cases. One of the considerations holding back the deployment of the technology is the concern of a court ruling that using copyrighted materials to train AI systems constitutes copyright infringement. Another consideration is that AI-generated works are not eligible for copyright protection.
As such, the legal stakes are high for both sides of the dispute.
