The Northern District of California’s District Court issued an order, on October 30, in the class action lawsuit initiated by artists Sarah Andersen, Kelly McKernan, and Karla Ortiz, which has brought into question the legal boundaries of artificial intelligence and copyright law. As already outlined HERE, at the heart of this dispute are AI software products that the plaintiffs argue were fed their creative works without consent, enabling them to generate new images mirroring the artists’ unique styles. The accusation leveled against Stability AI Ltd., DeviantArt, Inc., and Midjourney, Inc. is that they have, through their association with the AI tools, violated copyright by training the software with the plaintiffs’ art without authorization.
In reviewing the case, the court pointed out multiple areas where the plaintiffs’ legal argumentation was lacking. Primarily, the court underscored the absence of any claims that the AI-generated art bore a substantial likeness to the original copyrighted material of the plaintiffs, a key element in proving copyright infringement.
As the legal proceedings unfolded, each defendant responded with a motion to dismiss the allegations. Stability AI Ltd. rejected the claims on the grounds that there was no substantial similarity established between the AI-generated images and the plaintiffs’ works. DeviantArt, Inc. relied on the protections afforded by the Communications Decency Act, asserting that the plaintiffs did not convincingly argue the company’s contributory or vicarious liability in copyright infringement. Meanwhile, Midjourney, Inc. defended itself by highlighting the plaintiffs’ failure to demonstrate the company’s direct involvement with the contested AI system.
The court’s responses to these motions were varied. Stability AI Ltd. and DeviantArt, Inc. were both granted dismissals but with an opportunity for the plaintiffs to amend their complaints. Specifically, the court suggested that if the plaintiffs could demonstrate the AI’s output shared substantial similarities with their original works, or in DeviantArt’s case, if they could prove the company had an active role in the infringement, the claims might hold more weight. In contrast, the dismissal granted to Midjourney, Inc. was conclusive, with the court finding no evidence of the company’s engagement with Stable Diffusion in any capacity.
Additionally, the court has put on hold a decision regarding DeviantArt’s separate motion to strike down the right to publicity claim, which has been dismissed with a chance for revision. The court has expressed that the ruling on this motion will be contingent upon the re-presented claims of the plaintiffs, which need to navigate the preemptive scope of the Copyright Act and offer a substantial argument for the right to publicity claim.
This unfolding legal drama underscores the complexities that emerge at the intersection of AI technology and copyright law, setting a precedent for how similar cases may be adjudicated in the future. The outcome of this case could potentially reshape the way artists’ rights are protected in the era of machine learning and content generation.
Header image generated with Microsoft Bing; prompt: “the court underscored the absence of any claims that the AI-generated art bore a substantial likeness to the original copyrighted material of the plaintiffs”
