AI Art “In the Style of” & Contributory Liability

By: Jacob Alhadeff

Greg Rutkowski illustrates fantastical images for games such as Dungeons & Dragons and Magic the Gathering. Rutkowski’s name has been used thousands of times in generative art platforms, such as Stable Diffusion and Dall-E, flooding the internet with thousands of works in his style. For example, type in “Wizard with sword and a glowing orb of magic fire fights a fierce dragon Greg Rutkowski,” and Stable Diffusion will output something similar to Rutkowski’s actual work. Rutkowski is now reasonably concerned that his work will be drowned out by these hundreds of thousands of emulations, ultimately preventing customers from being able to find his work online. 

A picture containing nature

Description automatically generated

Examples of images generated by Dream Studio (Stable Diffusion) in Rutkowski’s style.

These machine learning algorithms are trained using freely available information, which is largely a good thing. However, it may feel unfair that an artist’s copyrighted images are freely copied to train their potential replacement. Ultimately, nothing these algorithms or their owners are doing is copyright infringement, and there are many good reasons for this. However, in certain exceptional circumstances, like Rutkowski’s, it may seem like copyright laws insufficiently protect human creation and unreasonably prioritizes computer generation.
A primary reason why Rutkowski has no legal recourse is because an entity that trains its AI on Rutkowski’s copyrighted work is not the person generating the emulating art. Instead, thousands of end-users are collectively causing Rutkowski harm. Since distinct entities cause aggregate harm, there is no infringement. By contrast, if Stable Diffusion verbatim copied Rutkowski’s work to train their AI before generating hundreds of thousands of look-a-likes, this would likely be an unfair infringement. Understanding the importance of this separation is best seen through understanding the process of text-to-art generation and analyzing each person’s role in the process. 

Text-to-Image Copyright AnalysisDiagram, text

Description automatically generated

To give a brief summary of this process, billions of original human artists throughout history have created art that has been posted online. Then a group like Common Crawl scrapes those billions of images and their textual pairs from billions of web pages for public use. Later, a non-profit such as LAION creates a massive dataset that includes internet indexes and similarity scores between text and images. Subsequently, a company such as Stable Diffusion trains its text-to-art AI generator on these text-image pairs. Notably, when a text-to-art generator uses the LAION database, they are not necessarily downloading the images themselves to train their AI. Finally, when the end user goes to Dream Studio and types in the phrase “a mouse in the style of Walt Disney,” the AI generates unique images of Mickey Mouse. 

A picture containing doll

Description automatically generated

A picture containing indoor

Description automatically generated
Examples of images generated by Dream Studio (Stable Diffusion) using the phrase “a mouse in the style of Walt Disney”

These several distributed roles complicate our copyright analysis, but for now, we will limit our discussion of copyright liability to three primary entities: (1) the original artist, (2) the Text-to-Image AI Company, and (3) the end-user. 

The Text-to-Image Company likely has copied Rutkowski’s work. If the Text-to-Image company actually downloads the images from the dataset to train its AI, then there is verbatim intermediate copying of potentially billions of copyrightable images. However, this is likely fair use because the generative AI provides what the court would consider a public benefit and has transformed the purpose and character of the original art. This reasoning is demonstrated by Kelly v. Arriba, where an image search’s use of thumbnail images was determined to be transformative and fair partly because of the public benefit provided by the ability to search images and the transformed purpose for that art, searching versus viewing. Here, the purpose of the original art was to be viewed by humans, and the Text-to-Image AI Company has transformatively used the art to be “read” by machines to train an AI. The public benefit of text-to-art AI is the ability to create complex and novel art by simply typing a few words into a prompt. It is more likely that the Generative AI’s use is fair because the public does not see these downloaded images, which means that they have not directly impacted the market for the copyrighted originals. 

The individual end-user is any person that prompts the AI to generate hundreds of thousands of works “in the style of Greg Rutkowski.” However, the end-user has not copied Rutkowski’s art because copyright’s idea-expression distinction means that Rutkowski’s style is not copyrightable. The end-user simply typed 10 words into Stable Diffusion’s UI. While the images of wizards fighting dragons may seem similar to Rutkowski’s work, they may not be substantially similar enough to be deemed infringing copies. Therefore, the end-user similarly didn’t unfairly infringe on Rutkowski’s copyright.

Secondary Liability & AI Copyright

Generative AI portends dramatic social and economic change for many, and copyright will necessarily respond to these changes. Copyright could change to protect Rutkowski in different ways, but many of these potential changes would result in either a complete overhaul of copyright law or the functional elimination of generative art, neither of which is desirable. One minor alteration that could give Rutkowski, and other artists like him, slightly more protection is a creative expansion of contributory liability in copyright. One infringes contributorily by intentionally inducing or encouraging direct infringement.

Dall-E has actively encouraged end-users to generate art “in the style of” artists. So not only are these text-to-art AI companies verbatim copying artists’ works, but they are then also encouraging users to emulate the artists’ work. At present, this is not considered contributory liability and is frequently innocuous. Style is not copyrightable because ideas are not copyrightable, which is a good thing for artistic freedom and creation. So, while the work of these artists is not being directly copied by end-users when Dall-E encourages users to flood the internet with AI art in Rutkowski’s style, it feels like copyright law should offer Rutkowski slightly more protection.

A picture containing text

Description automatically generated
An astronaut riding a horse in the style of Andy Warhol.
A painting of a fox in the style of Claude Monet.

Contributory liability could offer this modicum of protection if, and only if, it expanded to include circumstances where the copying fairly occurred by the contributor, but not the thousands of end-users. As previously stated, the end-users are not directly infringing Rutkowski’s copyright, so under current law, Dall-E has not contributorily copied. However, there has never been a contributory copyright case such as this one, where the contributing entity themselves verbatim copied the copyrighted work, albeit fairly, but the end user did not. As such, copyright’s flexibility and policy-oriented nature could permit a unique carveout for such protection.

Analyzing the potential contributory liability of Dall-E is more complicated than it sounds, particularly because of the quintessential modern contributory liability case, MGM v. Grokster, which involved intentionally instructing users on how to file-share millions of songs. Moreover, Sony v. Universal would rightfully protect Dall-E generally as due to many similarities between the two situations. In that case, the court found Sony not liable for copyright infringement for the sale of VHS recorders which facilitated direct copying of TV programming because the technology had “commercially significant non-infringing uses.” Finally, regardless of Rutkowski’s theoretical likelihood of success, if contributory liability were expanded in this way, then it would at least stop companies such as Dall-E from advertising the fact that their generations are a great way to emulate, or copy, an artist’s work that they themselves initially copied. 

This article has been premised on the idea that the end-users aren’t copying, but what if they are? It is clear that Rutkowski’s work was not directly infringed by the wizard fighting the dragon, but what about “a mouse in the style of Walt Disney?” How about “a yellow cartoon bear with a red shirt” or “a yellow bear in the style of A. A. Milne?” How similar does an end-user’s generation need to be for Disney to sue over an end-user’s direct infringement? What if there were hundreds of thousands of unique AI-generated Mickey Mouse emulations flooding the internet, and Twitter trolls were harassing Disney instead of Rutkowski? Of course, each individual generation would require an individual infringement analysis. Maybe the “yellow cartoon bear with a red shirt” is not substantially similar to Winnie the Pooh, but the “mouse in the style of Walt Disney” could be. These determinations would impact a generative AI’s potential contributory liability in such a claim. Whatever copyright judges and lawmakers decide, the law will need to find creative solutions that carefully balance the interests of artists and technological innovation. 

A picture containing doll

Description automatically generatedA yellow stuffed animal

Description automatically generated with low confidenceA picture containing text, fabric

Description automatically generated

The (Purple) Reign of Fair Use: Certiorari Granted for Warhol’s Portrait of Prince

By: Erika Hammer

Amongst several IP-focused cases this year having requested review by the Supreme Court, the high court has recently granted certiorari for a case involving copyright, fair use, and some famous individuals: artist Andy Warhol and musician Prince. The case focuses on whether a work is “transformative” under fair use, a major defense to copyright infringement. Notably, fair use is considered to be one of the most important exceptions to copyright law’s general monopoly grant of intelelctual property rights to authors of original works, as well as a major cornerstone for promoting artistic expression, access to knowledge, and dissemination of ideas. 

The case, Andy Warhol Foundation v. Goldsmith, arises from a set of portraits created by Andy Warhol, whose pieces often draw from preexisting works (e.g., a Marilyn Monroe photograph or a can of Campbell’s soup). The specific pieces at issue are portraits of Prince based on a Vanity Fair photograph taken by Lynn Goldsmith. The petition for certiorari describes how Warhol, via silkscreen printing, “cropped the image to remove Prince’s torso, resized it, altered the angle of Prince’s face, and changed tones, lighting, and detail” as well as “added layers of bright and unnatural colors, conspicuous hand-drawn outlines and line screens, and stark black shading that exaggerated Prince’s features.” 

Post-litigation, the district court granted the Andy Warhol Foundation summary judgment in favor of its fair use defense, deeming the use “transformative” for communicating a different meaning and message from the original Goldsmith work. However, the Second Circuit reversed, despite acknowledging that the two artists’ pieces represented different messages. It stated, “while the cumulative effect of those alterations may change the Goldsmith Photograph in ways that give a different impression of its subject, the Goldsmith Photograph remains the recognizable foundation upon which the Prince Series is built.” 

As the Andy Warhol Foundation argued in its petition for certiorari, the Second Circuit’s analysis focuses on the visual resemblances between the works. The Foundation further opines that this decision is creating a circuit split and highlights that the Ninth Circuit has held that a work of art is “transformative” when it portrays a different meaning or message from the original source. 

This case is significant not only because of the famous individuals involved, but also because it involves one of the most crucial doctrines in modern copyright law. Fair use, which is set forth in 17 U.S.C § 107, is the most wide-ranging limitation on copyright protection that attempts to promote the expression of artistic works. Fair use is also grounded in the goals of promoting common culture and enabling technological advancement. As such, highly creative works like Andy Warhol’s would appear to be exactly the kind of follow-on creativity that fair use is intended to not only protect, but to promote. 

Even if a work is highly creative, it must be examined under four factors used in determining whether there is a qualified fair use defense. These four factors include: (1) the purpose and character of the use, including whether the use is of a commercial nature or is for nonprofit, educational purposes; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and (4) the effect of the use upon the potential market. The Second Circuit found that each of these factors weighed in favor of Goldsmith.

Under the first factor, the more transformative a use is, the more likely said use is deemed to be fair. A foundational fair use case, Campbell v. Acuff Rose, emphasized transformative use as a critical factor. Transformative use is often seen as adding new, creative expression or changing the purpose or character of the copyrighted work. The more transformative a use is, the less significant the other fair use factors will be in the analysis. Typically, if a court finds transformative use under the first factor, that factor tends to strongly influence the inquiry into the rest of the fair use factors. 

With transformative use being such a crucial factor in fair use, which is of itself a crucial doctrine in copyright law, it comes as no surprise that the Supreme Court granted certiorari in this case. Despite the fact that Andy Warhol’s artwork appears to transform Prince’s depiction “from a vulnerable, uncomfortable person to an iconic, larger-than-life figure,” as described by the district court, this paradigm of the transformative nature of the work did not pass muster in the Second Circuit. 

In contrast, other prior Second Circuit cases that have been seminal in the “transformative” aspect of fair use have allowed use of the defense even when the original work is still a “recognizable foundation” to the subsequent piece at issue. Graham v. Dorling Kindersley held that a Grateful Dead biography that used copyrighted, original posters was fair use because they served a different purpose, despite the entirety of the original work being used in the follow-on biography. However, in opposition to certiorari, Goldsmith argues that the Warhol silkscreens shared the same purpose as Goldsmith’s copyrighted photograph, as well as the same essential artistic elements.

How the Supreme Court comes out on this decision – whether a different message or meaning is sufficient for transformation under fair use despite facial similarities – will be very important in copyright jurisprudence and the scope of fair use. 

An Infinity War: Will the Latest Marvel Lawsuits Finally Declare a Winner in the Battle of Copyright Termination Tests?

By Gracie Loesser

The Marvel Cinematic Universe is the highest-grossing franchise of all time, with a total worldwide gross profit of over $22 billion, and the demand for superhero content shows no signs of stopping.

So, it should not be a surprise that the original creators of characters like Iron Man, Spider-Man, and Dr. Strange are hoping to share in the success. Just this year, five former Marvel artists and their heirs have issued copyright termination notices, asserting their right to share in the Marvel Universe profits. In response, Disney-owned Marvel has sued, effectively asking the court to dismiss the creators’ ownership rights.

Whether the original Marvel artists will prevail depends on whether the creations are considered “works made for hire” under the federal Copyright Act. But how that will be determined is surprisingly unclear.

In 1976, Congress amended the Copyright Act addressing concerns that artists were losing valuable ownership rights through coercive agreements. The new Act incorporated a termination provision, which gave artists the right to reclaim any copyright interest they previously transferred regardless of any preexisting contract. Notably, the termination provision excluded any “work[s] made for hire,” with the rationale that an employer is considered the author and original copyright holder of any work made by an employee within the scope of their job. The issue is the law does not provide a definition of an “employer” and a “work made for hire” within the scope of the Act.

Courts have since stepped in to fill this gap. The Second Circuit developed the “instance and expense” test, which states that a work is made for hire “when the employer induces the creation of the work and has the right to direct and supervise the manner in which the work is carried out.” Under this inquiry, an independent contractor would be considered the author and original copyright holder of her work. However, the independent contractor could impliedly transfer her rights to the hiring party if certain conditions were met. The Second Circuit’s test proved popular, with the Fourth and Seventh Circuits adopting the approach.

This approach soon faced an existential challenge. In 1989, the Supreme Court attacked the appropriateness of the “instance and expense” test in CCNV v. Reid. The Court held that the Second Circuit’s analysis was not “consistent with the text of the Act,” stating the structure of Copyright Act § 101 necessitated two separate analyses: one for employees and one for independent contractors. This finding directly challenged the Second Circuit’s single test based on project control. To determine whether an artist should be classified as an employee or an independent contractor, the Supreme Court identified thirteen non-exhaustive factors rooted in the common law of agency to guide the analysis.

Although the Supreme Court’s ruling appeared to rebuke the “instance and expense” analysis, the Second Circuit continued to apply the test in the years following the Reid decision. In Marvel Characters, Inc. v. Kirby, the Second Circuit was asked to review the district court’s grant of summary judgment in favor of Marvel. The lower court had applied the “instance and expense” test and determined that former Marvel artist Jack Kirby’s drawings were works made for hire, thus invalidating the termination notices issued by his heirs. Upon review, the Second Circuit affirmed the district court’s use of the “instance and expense” test, acknowledging and simultaneously dismissing the Supreme Court’s criticism in a brief footnote. Kirby appealed the decision hoping that the Supreme Court would grant certiorari to once and for all dismiss the “instance and expense” test. Kirby’s cause was given further support when the former director of the U.S. Patent and Trademark Office, Bruce Lehman, filed a brief urging the Court to use this opportunity to reject the “instance and expense” analysis as improper. Unfortunately, Marvel settled the case, just as it looked like the Kirby campaign might succeed in convincing the Supreme Court to grant certiorari.

Since the ruling in Kirby, the Second Circuit’s rulings have only added to the confusion. In Horror Inc v. Miller, the court was again asked to analyze the validity of an author’s copyright termination. In that case, writer Victor Miller contended he was an independent contractor under the Copyright Act when he created the original Friday the 13th screenplay, making his termination notice enforceable. Surprisingly, the Second Circuit agreed, relying on the Supreme Court’s thirteen Reid factors in determining that Miller’s screenplay was not a work made for hire. Despite this major shift from the Circuit’s previous defense of their “instance and expense” approach, the court did not acknowledge the change or explain their reasoning.

Now, the Second Circuit is faced with five potential Marvel copyright termination cases, all of which closely resemble the facts in Kirby. Assuming the Second Circuit continues to apply the Reid analysis as it did in Miller, the outcome will likely be in the artists’ favor.

However, that assumes the Circuit maintains this approach. Based on the content of at least one complaint, Marvel’s counsel may try to convince the court to reverse course. The filed claims against the company’s former artists notably include separate considerations of “instance” and “expense.” Given the Circuit’s inconsistent rulings in the last decade, the outcome of these latest cases is less than certain, making this an important opportunity for the court to clarify its stance. Nevertheless, even if the Second Circuit decides to abandon the “instance and expense” test, whether the Fourth and Seventh Circuits will make similar changes to their analysis remains to be seen.

Is Code Killing Copyright?

By: Katherine Czubakowski

Early last month, the Supreme Court released its long-awaited decision in Google LLC v. Oracle America Inc.  The Court found that Google’s unauthorized copying of 11,500 lines of code from Oracle’s Java SE API was fair use because Google took only as much code as it needed to create a new and transformative program. While some argue that this outcome protects fundamental aspects of how code is created and the technology industry, others see this decision as a significant blow to copyright protections. This disagreement comes down to a fundamental question the Supreme Court seems to have side-stepped in this case: whether code should be protected under copyright at all.

An API, or Application Programming Interface, is a list of actions one can take regarding specific software and how one would take those actions. For example, if gardening were a software, you could choose the action you want to perform (dig, e.g.) and how you want to perform that action (with a shovel, a hoe, a pickaxe, your hands, etc.). The Java API in question contains a basic list of common actions (sorting a list, for example) and how those actions are accomplished (alphabetically, numerically, etc.). When Google began developing the Android software used in their smartphones, they wrote their own code to tell the program what to do and how to do it, but copied the declaring code—the part of the program which matches the name assigned to each task with the program necessary to perform the task—from 37 of Java’s listed tasks. By doing so, the programmers working on the Android software were able to continue using the commands with which they were familiar, such as PrintLn() (which tells the program to print the specified text on the user’s screen) and (which tells the computer to display the user’s current date and time), in their own code, but these commands relied on Google’s newly written code to perform the task.

In determining that Google was legally allowed to copy this code, the Court relied on the doctrine of fair use.  Although copyright owners generally hold exclusive rights to create derivative works, which are new works based on their own pre-existing work, fair use is a legal exemption which allows someone to use copyright protected work without the author’s permission in certain circumstances. Courts consider fair use on a case-by-case basis and analyze four different aspects of the otherwise-infringing use: its purpose and character, the nature of the copyrighted work, the amount and substantiality of the portion used, and the effect on the potential market for the copyrighted work.  In its recent cases regarding fair use, the Court has created a sub-factor that it considers under the purpose and character of the use: transformativeness. A work is considered transformative when it uses the original copyrighted work in an unexpected way or in a way which alters the original meaning or message.  Transformativeness weighs strongly in favor of fair use because it encourages creativity and furthering of the arts.  This sub-factor frequently affects all four factors in the fair use analysis and can sometimes even outweigh the importance of the other three factors. It can often be difficult to tell if a work used in a transformative way is a derivative work or if it falls under the fair use exception.

In Google LLC v. Oracle America Inc. the Court’s decision hinged on its finding that Google’s use was transformative. The Court first analyzed the nature of the work and found that APIs were fundamentally different than other types of code. Because the declaring code fuses together the uncopyrightable idea of how the code is organized with the copyrightable code which tells the computer how to perform a function, the Court saw the copied code as valuable only as a result of the programmer’s investment in learning it. Since the copied code did not hold independent value, the Court felt that applying fair use in this circumstance would not undermine general copyright protection for other programs. The Court then turned to the purpose and character of the use, which is where they discussed the work’s transformative nature. It found that Google’s purpose in using the copied code was “to create a different task-related system for a different computing environment” than the creators of Java had originally intended.  Google’s use of the code was part of the “creative progress” which the Court saw as copyright law’s objective, so they found that the use was transformative. The Court further found that, although Google copied “virtually all of the declaring code needed to call up hundreds of different tasks,” they copied a relatively small amount of the total API in question. Because this relatively small portion of the API was tied to a valid and transformative purpose, the Court felt that the third factor weighed in favor of fair use as well. Finally, the Court found that Android was not a market substitute for Java SE because the two products were substantially different. Weighing all these factors together, the Court found that because they only took as much as was necessary to allow their programmers to use “accrued talents to work in a new and transformative program,” Google’s “reimplementation of a user interface” was protected by the fair use doctrine.

The Court’s analysis and reliance on transformation in this case presents a danger to those seeking to copyright their code. This is because code is fundamentally different than many other works protected by copyright; it combines functionality with creative expression. Unlike traditionally copyrightable works, programs are usually created in a way which relies on previously created code to function. When writing new code, very few programmers actually write code which can interact directly with the computer. Instead, they use one of a number of programs which translate a more readable code, such as Java, into code which the computer can understand. Without being able to copy some fundamental aspects of the language, programmers would have to create a new language anytime they wanted to write new code. In practice, this means that many different programs with different purposes all rely on the same underlying program(s) to translate their code into a form the computer can understand. 

Although the Court likely reached the correct outcome in this case, the repercussions of its decision in other fields damages traditional copyright holder’s rights. The Court’s transformative analysis fails when applied in the context of programming because a program’s reliance on other code is a necessary aspect of its creation. Thousands of substantially different programs rely on the same underlying code in order to function. However, purely creative expression does not have this same reliance on preexisting works—as evidenced by Congress’s grant of derivative works rights to copyright holders. By trying to fit both pure creative expression and functional creative expression under the same body of law, the Court has blurred the lines between what is transformative and what is derivative and has put at risk the exclusive rights guaranteed to copyright owners of traditionally copyrightable works.

Under the Sea, Under the ©? Ninth Circuit Considers Art Inspired by Nature

Folkens (left) vs. Wyland (right)

By Yonah Reback

The intersection of copyright law and the animal kingdom reached new heights during the global phenomenon of Naruto v. Slater. When wildlife photographer David Slater traveled to the jungles of Indonesia in 2011, he surely expected to take snapshots of its native inhabitants. To Slater’s surprise, a macaque nicknamed Naruto grabbed his camera and took what soon became the infamous “monkey selfie.” Thus began a six-year legal saga as to whether Naruto or Slater owned the copyright for the photograph. Although the U.S. Copyright Office indicated that there was no basis for extending copyright ownership to animals, the case—filed by PETA in defense of the monkey—ultimately settled, leaving the public wanting for legal analysis of the issue.

Continue reading