Artificial intelligence (AI), when used in art, faces several legal challenges. The most recent is a class action lawsuit filed by artists Sarah Andersen, Kelly McKernan, and Karli Ortiz, who accused the London-based Stability AI Ltd, and its US affiliate Stability AI, Inc, of a variety of crimes.
The lawsuit was filed on 13th January in the US District Court of San Francisco. It accuses Stability AI of creating AI software, Stable Diffusion, that downloaded millions of images protected by copyright from the internet. They then stored them in compressed copies (or “diffused”) and made them available to Midjourney and Deviant Art. Both companies are named in the suit. The complaint states that the “complex collage” tool of the software allows users to alter the original images and create adapted versions that can be sold. The complaint claims that “these ‘new images’ are based on” copyrighted pictures found on the internet “and are derivatives.”
According to the US Copyright Office, “Only the copyright owner in work can prepare or authorize another to create an adaptation of that piece,” In any case, where a copyrighted piece of work is used illegally without the owner’s permission, the copyright protection does not apply to the part of the work that has been unlawfully used. Therefore, unauthorized adaptations of works may be considered copyright infringement.
Matthew Butterick is one of the attorneys representing the artists involved in the class action lawsuit. He claims that derivative works created using copyrighted pictures “may or may” not resemble the originals. “However, they are copies of [underlying] images and compete against them in the market.” Stable Diffusion’s ability to flood markets with an almost unlimited number of infringing images will cause permanent damage to the art market and artists.
Getty Images has filed a separate copyright lawsuit in London against Stability AI. Getty Images is a visual media company that provides stock images, editorial photographs, videos, and music to businesses and consumers. It has more than 477,000,000 assets. Getty Media’s spokesperson claims that Stability AI chose to ignore long-standing legal rights and viable licensing options in order to pursue its own commercial interests. Getty seeks damages of $2 trillion each in the UK and USA for this alleged infringement.
Stability AI’s spokesperson says, “We take this matter seriously.” However, allegations are based on a misunderstanding of how Stability AI’s technology and law work.
Megan Noh, a New York-based lawyer, says that the issue of what is and isn’t allowed in reusing copyrighted materials is still not completely settled. The “fair use” exception to copyright infringement is “a developing area,” with AI applications giving new life to the multi-factor test of [copyright law]. This includes the question of “transformativeness” (which is still in flux, pending the Supreme Court ruling in [Andy] Warhol Foundation’s v [Lynn] Goldsmith’s case) and the impact of AI-generated art on the market for the original works.
Creation vs. generation
US Copyright Office is the only one to have made a firm decision on this issue, refusing to register copyrights for pieces created–the word “created,” however, was not used—-through AI. Joshua Kaufman is a partner at the Washington-based Venable law firm. He says that those who create AI programs to produce artwork have asked the Copyright Office to reconsider this issue. According to Kaufman, the current law makes it difficult for developers to protect their AI-generated images from those who would copy them. He says that AI creators that want copyright protection argue that because corporations can own copies and are not humans, they should be considered creators under the doctrine of work-for-hire. Therefore, AI-created artworks could be given the same protection.
AI is the current Rorschach Test in the art world. Some see only landmines, while others see opportunity. Amelia K. Brankov, a New York attorney, says that there are concerns “that AI-generated art could impact the livelihood of artists like illustrators and graphics artists who rely on commissions. Their clients may want to cut costs by using an AI image generator instead of commissioning a new piece of work from an artist.” Some artists, on the other hand, are embracing technology and incorporating it into their practice. They are also trying to understand the rights and obligations they have with regard to AI-generated output.