Generative AI Lawsuit Takes Center Stage in Courts

A high-profile generative AI lawsuit has made its way to the courts, raising important questions about the ownership and control of AI-generated content. This lawsuit centers around a dispute over the rights to a particular type of AI-generated image.

The lawsuit claims that the defendant used the plaintiff's AI model to create images that were then sold for profit, without permission or compensation. The plaintiff is seeking damages and a court order to stop the defendant from using their AI model.

The case highlights the need for clear guidelines and regulations around the use of generative AI technology. As AI-generated content becomes increasingly prevalent, it's essential that we establish a framework for ownership and control.

Related reading: Generative Ai Content

Getty has brought two lawsuits against Stability, one in the U.S. and one in the U.K., claiming Stability copied more than 12 million photographs from Getty websites to train its Stable Diffusion image model.

A fresh viewpoint: One Challenge in Ensuring Fairness in Generative Ai

Getty claims the pairing of high-quality photographs with concise descriptions makes the images more valuable to Stability as training data than ordinary uncaptioned images available on the Internet.

Getty owns copyright on many of these photographs and holds nonexclusive licenses on many others, which allows them to sue Stability for infringement of copyrights and seek remedies against it for images in which Getty owns rights.

In the Andersen case, the named plaintiffs claim to represent a class of all visual artists whose copyrights Stability has infringed, and argue that all of Stable Diffusion's outputs are infringing derivative works, not just those that are substantially similar to the originals.

Sarah Andersen, a graphic artist, is the lead plaintiff in a proposed class action lawsuit against Stability, similar to the one brought by Getty.

The complaint alleges that Stability used images and captions from Getty websites as training data for its image model, Stable Diffusion.

If this caught your attention, see: Generative Ai by Getty

Stability is accused of copying more than 12 million photographs, along with textual descriptions of the photographs' contents, from Getty websites.

Getty claims that Stability used these high-quality images and concise descriptions to make its image model more valuable.

The complaint states that the outputs produced by Stable Diffusion are often substantially similar to, and therefore infringing derivatives of, images copied from the Getty websites.

Andersen also charges Stability with violating a law that protects copyright management information from being removed or altered.

The complaint shows examples of mangled Getty Images logos on Stable Diffusion outputs that are similar in content to those embedded in Getty photos.

Andersen claims to represent a class of all visual artists whose copyrights Stability has infringed, and asserts that all of Stable Diffusion's outputs are infringing derivative works.

You might enjoy: Explainable Ai Generative Diffusion Models

Thomson Reuters v. Ross Intelligence is a significant lawsuit involving generative AI. It was filed in the U.S. District Court for the District of Delaware in May 2020.

The case centers on allegations that copyrighted headnotes from Thomson's Westlaw legal research database were used as training data for an AI legal research tool developed by ROSS Intelligence.

In September 2023, the court ruled that most issues in the case could not be resolved on summary judgment. Many of the critical facts in the case remain genuinely disputed.

The court did, however, grant the plaintiffs' motion for summary judgment on the question of whether ROSS engaged in at least some actual copying of materials from Westlaw.

Among the key facts the court noted were in dispute related to fair use is the aims and outcomes of the training process. The court did not find that intermediate copying made in the training process is always transformative for purposes of evaluating the first fair use factor.

The court said that whether the intermediate copying is transformative depends on the precise nature of the use. If the AI tool only studied the language patterns in the headnotes to learn how to produce judicial opinion quotes, then it would be transformative intermediate copying.

A different take: Generative Ai Legal Issues

But if ROSS used the untransformed text of headnotes to get its AI to "replicate and reproduce the creative drafting done by Westlaw's lawyer-editors", then those intermediate copying cases would not apply.

The question of whether this use was transformative is a material question of fact that the jury needs to decide. The court noted that "the first fair use factor comes down to the jury's finding of transformativeness."

Factual disputes also remain for trial regarding the remaining fair use factors. For the second factor, disputes about how original the copied Westlaw material is must go to a jury.

The court found that disputes about whether the material is in fact protected, and how far that protection extends, must also go to a jury.

For the third factor, substantiality of the use, a jury must resolve disputes related to whether the scale of copying was practically necessary and furthered its transformative goals.

And for the fourth factor, potential marketing impact, a jury must ultimately decide hotly debated questions related to market harm.

Explore further: Westlaw Generative Ai

The outcome of Getty's claim against Stability, and other similar cases, will have far-reaching implications for the generative AI industry.

Developers may be able to train their AI solutions using copyright protected material outside of the UK, where they are subject to less restrictive AI regulation or even legislative exemptions.

The case against ROSS Intelligence Inc. is expected to go to trial in August 2024, with many of the critical facts still in dispute. The court found that intermediate copying made in the training process is not always transformative for purposes of evaluating the first fair use factor.

A jury will ultimately decide the question of whether the use was transformative, which will be central to the first fair use factor. The court also found disputes about how original the copied Westlaw material is and how far that protection extends.

The outcome of this case may have wide-ranging consequences for the industry, including the possibility of developers "jurisdiction shopping" to train AI models in more favourable legislative landscapes.

Developers seeking to publish games on Steam are now required to disclose when a game contains pre-generated AI content and promise that it is not 'illegal or infringing'.

For more insights, see: Impact of Generative Ai on Tax Industry

Copyright infringement claims against generative AI systems are complex and multifaceted.

Courts have repeatedly held that to infringe the derivative work right, copyright owners must show substantial similarity between the second work's expressive elements and the first work's expression so it is fair to infer the defendants improperly appropriated expressive elements from the copyright owners' works.

Plaintiffs may try to focus on the open-ended part of the derivative work definition, but most judicial decisions have not given a broad interpretation to that phrase.

Developers of generative AI systems can guard against derivative work infringement by removing duplicate images and/or developing output filters to prevent the resulting image from infringing.

A key issue in the Stability cases is whether the outputs of generative AI systems are copyrightable.

The training and development claim in the Getty vs. Stability case centered on whether Stability's use of Getty's images to train and develop Stable Diffusion amounts to copyright infringement under the Copyright Designs and Patents Act 1988 (CDPA) in the UK.

If this caught your attention, see: How Are Modern Generative Ai Systems Improving User Interaction

Stability argued that no training or development of Stable Diffusion took place in the UK, but the judge was not convinced and refused to throw out Getty's claim.

The secondary infringement of copyright claim alleges that Stability has unlawfully imported into, possessed and/or dealt with an 'article', namely the pre-trained Stable Diffusion software, in the UK.

The judge refused to throw out this claim, citing the need to determine the true interpretation of the word 'article' in sections 22, 23 and 27 of the CDPA.

Here are the key sections of the CDPA relevant to the secondary infringement claim:

The training and development of Stable Diffusion was a key issue in the lawsuit between Getty and Stability AI.

Mrs. Justice Smith was not convinced by Stability's claim that no training or development of Stable Diffusion took place in the UK, despite the company's evidence suggesting otherwise.

Stability's founder, Mr. Emad Mostaque, testified that no Stability employee based in the UK worked on developing or training Stable Diffusion.

For another approach, see: Learning Generative Ai

However, Mrs. Justice Smith pointed out that Mr. Mostaque's previous media appearances and YouTube interviews seemed to contradict his evidence.

The court's decision was to send both Getty and Stability to the disclosure process, which will require them to reveal the precise nature of the development and training of Stable Diffusion.