If you were around in the 1980s, you will undoubtedly remember how exciting the promise of artificial intelligence, or AI was and how prevalent it was in pop culture. Who didn’t want a sentient, talking car like K.I.T.T. from the television show Knight Rider? Or a robot friend like Johnny Five from the movie Short Circuit? Blade Runner’s replicants, Daryl (an acronym for Data-Analyzing Robot Youth Lifeform) from the movie D.A.R.Y.L., and Lisa—created by two techno-savvy teenagers in the movie Weird Science—pushed AI beyond semiconductors and presented them in human form. And TRON gave us a passport to an incredible world within a CPU and was one of the first movies to use computer-generated imagery. It’s now the 2020s, and AI is finally here on an end-user level. One of its most visible uses is image generation from text prompts, not unlike the DOS interface of Ferris Bueller’s computer as he hacked his school’s mainframe from his bedroom; who could forget that famous line, “I asked for a car; I got a computer. How’s that for being born under a bad sign?” Its arrival triggered questions over copyright, some that remain open today.
Technology advances are independent of the law, but copyright law has been plenty flexible to cover advances as they occured. From the printing press through the Betamax, the U.S. Supreme Court observed, “[f]rom its beginning, the law of Copyright has developed in response to significant changes in technology. Indeed, it was the invention of a new form of copyright equipment—the printing press—that gave rise to the original need for copyright protection. The latest of the technological advances—artificial intelligence, or “AI”—gained widespread recognition in the news and caused an uproar in the arts community when an image titled “Théâtre D’opéra Spatial” was entered in the 2022 Colorado State Fair’s annual fine art competition and won. The author then challenged the U.S. Copyright Office after he was denied registration and lost.
One thing that has remained constant about copyright throughout the centuries and advances in technology is the requirement of human authorship. “Copyright is designed to adapt with the times. Underlying that adaptability, however, has been a consistent understanding that human creativity is the sine qua non of copyrightability, even as that human creativity is channeled through new tools or into new media.” In September this year, the U.S. Copyright Review Board and the U.S. District Court for the District of Columbia have both reaffirmed that understanding, stating that “[h]uman authorship is a bedrock requirement of copyright.”
This article examines the Copyright Act’s originality requirement and examines why a machine that has the capacity to create images according to its user’s input does not satisfy the human creativity requirement needed to accrue copyright protection.
Copyright protection subsists in “original works of authorship fixed in any tangible medium of expression[.]” Copyright registration is permissive, meaning that protection applies without registering a work with the U.S. Copyright Office, so long as the requirements of originality and fixation in a tangible medium of expression are met, as stated: “A valid copyright exists upon a qualifying work’s creation and ‘apart’ from registration, however; a certificate of registration merely confirms that the copyright has existed all along.” The fixation requirement is satisfied and copyright attaches the moment the paintbrush touches the canvas, the moment the shutter is pressed on the camera, and the moment CMD-S is pressed on the keyboard.
The originality requirement, characterized by the U.S. Supreme Court as “founded in the creative powers of the mind,” is a low threshold. It has been described by courts as a “modicum of creativity” and a “creative spark.” So long as the work is an independent creation and contains a minimal degree of creativity by the author, the originality requirement is satisfied.
AI image generators utilize artificial neural networks to create images based on text prompts from a human. “These technologies ‘train’ on vast quantities of preexisting human-authored works and use inferences from that training to generate new content. Some systems operate in response to a user’s textual instruction, called a ‘prompt.’” For example, when a user enters a text prompt, such as a “wide-angle shot from below of a female astronaut with an athletic female body striding across Mars in an infinite universe, synthesizer waves, digital art,” the generator pulls what it has “learned” from its “training” on perspective, females, astronauts, physiology, the planet Mars, and outer space and combines them to form an image. This was the exact text prompt that Cosmopolitan Magazine used to produce the first AI-generated magazine cover last year.
The fact that AI image generators are trained by referencing copyrighted images on the Internet raises copyright questions of whether such activity is infringement and, if so, whether fair use is a defense. In September 2022, Getty Images announced that it was banning AI-generated images from its library because of concerns over unsettled copyright questions. However, in March 2023, Getty Images announced a collaboration with computer graphics chip maker NVIDIA to train its Picasso AI model on Getty’s closed, fully licensed library in an effort to create “sterile” AI-generated images without exposure to unresolved copyright issues. Getty Images’ Chief Product Officer stated, “Both companies want to develop these tools in a responsible way that returns benefits to creators and doesn’t pass risks on to customers, and this collaboration is testament to the fact that’s possible[.]”
Although questions of infringement and whether fair use could be asserted as a defense are outside the scope of this article, it is important to acknowledge that they exist. Indeed, the U.S. Copyright Office issued a Notice of Inquiry on Artificial Intelligence and Copyright on August 30, 2023, seeking input on questions such as these.
Is creativity an exclusively human trait? Courts have already answered this existential question through over a century of case law as far as it pertains to copyright protection. The answer is for copyright protection to subsist, a work must be created by a human. “The 1976 [Copyright] Act’s ‘authorship’ requirement as presumptively being human rests on centuries of settled understanding.” The U.S. Copyright Office has stated that it “will register an original work of authorship, provided that the work was created by a human being[,]” and “the Office will refuse to register a claim if it determines that a human being did not create the work.” The U.S. Copyright Office’s Compendium of U.S. Copyright Practices explains, “the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author. The crucial question is ‘whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.’”
Burrow-Giles Lithographic Co. v. Sarony was the first case to address the issue of human creativity as a necessary element for copyright protection. New technology—the camera—made it possible to capture as an image what the human eye saw. The Court held the resulting works—photographs—were entitled to copyright protection because the camera was a tool and the photographer—selecting the subject and framing the shot—was the creator of the image: “We entertain no doubt that the constitution is broad enough to cover an act authorizing copyright of photographs, so far as they are representatives of original intellectual conceptions of the author.” The court explained: “The only reason why photographs were not included in the extended list in the [Copyright] act of 1802 is, probably, that they did not exist, as photography, as an art, was then unknown, and the scientific principle on which it rests, and the chemicals and machinery by which it is operated, have all been discovered long since that statute was enacted.” The Burrow-Giles court held for the first time that the utilization of new technology requires human creative input in order to accrue copyright protection:
[P]hotography is to be treated for the purposes of the act as an art, and the author is the man who really represents, creates, or gives effect to the idea, fancy, or imagination. . . . These views of the nature of authorship and of originality, intellectual creation, and right to protection, confirm what we have already said.
In a 1941 copyright infringement case, Oliver v. Saint Germain Foundation, the plaintiff published a book titled A Dweller on Two Planets that he claimed was dictated to him by the spirit of a deceased person. The defendant moved to dismiss for an invalid copyright registration. The District Court for the Southern District of California, Central Division, observed: “It is perfectly clear, therefore, that Oliver wished to impress in the strongest terms possible, his sincere belief in the truthfulness of his statement that he, a mortal being, was not the author, and to induce those who might read to believe that it was dictated by a superior spiritual being[.]” It held, “The law deals with realities and does not recognize communication with and the conveyances of legal rights by the spiritual world as the basis for its judgment. Nevertheless . . . [o]ne who narrates matters of fact may be protected by copyright as to his arrangement, manner and style, but not as to material or ideas therein set forth.” The case was dismissed.
Urantia Foundation v. Maaherra was a 1997 copyright infringement opinion where both parties believed “the Urantia Book was authored by celestial beings and transcribed, compiled and collected by mere mortals.” The defendant distributed a digitized version of the book on computer disk free of charge to other individuals, and when the Urantia Foundation sued, the defendant conceded copying of the book but maintained that the copyright registration was invalid. The Ninth Circuit Court of Appeals stated, “We agree with Maaherra, however, that it is not creations of divine beings that the copyright laws were intended to protect, and that in this case some element of human creativity must have occurred in order for the Book to be copyrightable. At the very least, for a worldly entity to be guilty of infringing a copyright, that entity must have copied something created by another worldly entity.” The case was reversed and remanded on grounds other than the originality requirement.
Naruto v. Slater was a 2018 copyright infringement opinion over picture selfies taken by Naruto, a crested macaque primate. David Slater was a photographer. Slater was in Indonesia and left his camera unattended. Naruto allegedly took several selfies with Slater’s camera. Slater published the selfies in a book. Wikimedia published one of the selfies, claiming it was not protected by copyright. People for the Ethical Treatment of Animals (PETA), claiming to represent Naruto, sued Slater for copyright infringement, claiming Naruto was the copyright claimant. The District Court for the Northern District of California held:
Here, the Copyright Act does not “plainly” extend the concept of authorship or statutory standing to animals. To the contrary, there is no mention of animals anywhere in the Act. The Supreme Court and Ninth Circuit have repeatedly referred to “persons” or “human beings” when analyzing authorship under the Act. … Despite Next Friends’ assertion that declining to grant a monkey copyright to a photograph “would depart from well-established norms,” Next Friends have not cited, and I have not found, a single case that expands the definition of authors to include animals.”
The Ninth Circuit Court of Appeals, viewing the appeal from the lens of standing, observed that only humans are contemplated under the Copyright Act: “The terms ‘children,’ ‘grandchildren,’ ‘legitimate,’ ‘widow,’ and ‘widower’ all imply humanity and necessarily exclude animals that do not marry and do not have heirs entitled to property by law.” The district court’s dismissal was affirmed.
It is an unsettled question whether generative AI constitutes true creativity. For example, a machine trained on what it has learned from images of dishes and cutlery will be able to generate an image from a prompt of “a spoon in a bowl” but may not be able to produce satisfactory results from a prompt of “a bowl on a spoon” because it cannot imagine what that would look like. Whether generative AI is genuinely creative is probably best left for philosophers and programmers to debate; however, the Thaler court (described infra) observed, “[t]he issue of whether non-human sentient beings may be covered by ‘person’ in the Copyright Act is only ‘fun conjecture for academics[.]’” Nearly sixty years ago, the Copyright Office was already acknowledging issues with computer-generated works of authorship and its 1965 annual report recognized recent advances in computer abilities and observed:
The crucial question appears to be whether the “work” is basically one of human authorship, with the computer merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.
This question has just recently been answered from the following two cases, both decided in the past month as of this writing, holding that if the creative input does not come from a human, then copyright protection does not accrue.
In August this year, the District Court for the District of Columbia dealt with what could be the very first AI-generated image submitted to the Copyright Office for registration in Thaler v. Perlmutter. The plaintiff, Thaler, owned a computer system he called the Creativity Machine that generated an image he titled “A Recent Entrance to Paradise,” which he then tried to register in 2018. “Thaler left a note for the Office stating that the Work ‘was autonomously created by a computer algorithm running on a machine’ and he was ‘seeking to register this computer-generated work as a work-for-hire to the owner of the Creativity Machine.’” When his second request for reconsideration was denied, Thaler sued.
The court distinguished the camera in Burrow-Giles Lithographic Co. v. Sarony with Thaler’s claim of autonomous creation by stating, “[h]uman involvement in, and ultimate creative control over, the work at issue was key to the conclusion that the new type of work, photographs, fell within the bounds of copyright.” The court acknowledged:
Undoubtedly, we are approaching new frontiers in copyright as artists put AI in their toolbox to be used in the generation of new visual and other artistic works. The increased attenuation of human creativity from the actual generation of the final work will prompt challenging questions regarding how much human input is necessary to qualify the user of an AI system as an “author” of a generated work, the scope of the protection obtained over the resultant image, how to assess the originality of AI-generated works where the systems may have been trained on unknown pre-existing works, how copyright might best be used to incentivize creative works involving AI, and more.
Although Thaler asserted new facts that he provided instructions and directed his AI to create the image and the AI was entirely controlled by him, he was estopped from claiming such because of his previous conflicting statements to the Copyright Office. The court then held, “this case presents only the question of whether a work generated autonomously by a computer system is eligible for copyright. In the absence of any human involvement in the creation of the work, the clear and straightforward answer is the one given by the Register: No.”
On September 5, 2023, the Copyright Review Board’s Second Request for Reconsideration for Refusal to Register Théâtre D’opéra Spatial denied copyright protection to the AI-generated image that won the 2022 Colorado State Fair’s annual fine art competition and garnered significant backlash from the arts community. Jason Allen created the image using Midjourney, a text-to-image AI program, but did not disclose it on his registration application. The Copyright Office knew about AI’s involvement from all of the publicity the image attracted and asked Allen about his use of Midjourney. Allen stated that he “input numerous revisions and text prompts at least 624 times to arrive at the initial version of the image.” He then used Adobe Photoshop to revise and add to the image and used Gigapixel AI to increase the resolution of the image. In response to his explanation, the Copyright Office asked him to exclude the content generated by Midjourney and he declined. The Copyright Office then refused to register the image. Allen requested reconsideration, stating that the point of creativity was his series of prompts to adjust the scene and dictate the tone of the image. He argued that the AI-generated image constituted raw material that he then transformed himself. The Office refused to register the image a second time, and he made a second request for reconsideration.
The Copyright Review Board was not persuaded by Allen’s arguments in his second request for reconsideration that his prompts were “the essential element of human creativity required to create a work using the Midjourney program.” However, the Board acknowledged that the prompts themselves could be afforded copyright protection as literary works if they were sufficiently creative. In denying his second request, the Copyright Review Board stated, inter alia: “Because Mr. Allen has refused to disclaim the material produced by AI, the Work cannot be registered as submitted.”
The Thaler court had held, “Copyright has never stretched so far, however, as to protect works generated by new forms of technology operating absent any guiding human hand, as plaintiff urges here. Human authorship is a bedrock requirement of copyright.” Why can’t, as both Allen and Thaler argued, the point of human creativity be the prompts fed to the machine? There are two reasons why. The first is there is longstanding precedent in analogous commission arrangements with artists for the prompts not to be the point of creativity. The second is the long-retired “sweat of the brow” doctrine, which prohibits copyright protection from accruing merely through one’s hard work.
Commission arrangements with artists are a direct analog to a user inputting text prompts into the machine to generate an image. With a commission arrangement, the commissioner hires the artist and tells the artist what (and sometimes how) to create. The artist may be given free reign or micromanaged by the commissioner, or anywhere in between. Likewise, when a user inputs prompts into the machine, the user gives as little or as many prompts as desired. When the artist shows the commissioner the finished product, the commissioner may decide it needs a little of this and a little of that, and directs the artist accordingly. Likewise, when the user receives an image from the machine, the user may decide to add to or refine the prompts given to adjust the image.
In a commission arrangement with an artist, the point of human creativity is not the commissioner’s directions but the artist interpreting the directions, no matter how exacting the commissioner’s directions may be. If a commissioner gives four artists the same directions, he will receive four different works of art in return because it is the artist’s personality, emotions, and style that are the point of creativity, not the commissioner’s directions. One hundred twenty years ago, the U.S. Supreme Court observed: “Personality always contains something unique. It expresses its singularity even in handwriting, and a very modest grade of art has in it something irreducible, which is one man’s alone. That something he may copyright unless there is a restriction in the words of the act.” Absent any prior written work-made-for-hire agreement with the artist, copyright protection vests with the artist when that brush touches the canvas because “copyright protection attaches immediately upon the creation of original works of authorship fixed in any tangible medium of expression[.]”
However, with a machine, “when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology—not the human user.” Similar to a commissioner giving four artists the same directions and receiving four different works in return, many AI image generators will return multiple images from the same text prompt. Indeed, the Copyright Office found the analog between machine and commissioned artist as the point of creativity valid. In guidance published in the Federal Register regarding works containing AI-generated content, the Copyright Office stated, “users do not exercise ultimate creative control over how such systems interpret prompts and generate material. Instead, these prompts function more like instructions to a commissioned artist—they identify what the prompter wishes to have depicted, but the machine determines how those instructions are implemented in its output. . . . When an AI technology determines the expressive elements of its output, the generated material is not the product of human authorship.” Accordingly, the point of creation for AI-generated images is not the text prompt.
Bare text prompts, without anything else, may not be sufficiently creative to accrue copyright protection as literary works. In addition to possibly being “words and short phrases” not protectible under 37 C.F.R. § 202.1(a), or “ideas, plans, methods, systems, or devices” not protectible under 37 C.F.R. § 202.1(b), such prompts could constitute unoriginal facts prevented from accruing copyright protection under the long-retired “sweat of the brow” doctrine. In Feist Publications, Inc. v. Rural Telephone Service Co., Inc., Feist Publications asked to license Rural Telephone Service Co.’s telephone directory information for Feist’s own directory, and when Rural Telephone Service Co. refused, Feist Publications used it anyway. The U.S. Supreme Court reaffirmed that the sweat of the brow doctrine was no longer valid by holding that information contained in a telephone directory constituted facts, and while Rural Telephone Service Co. spent a lot of time and effort gathering the information for its telephone directory, which Feist Publications then used, that information was not protected by copyright.
The sweat of the brow doctrine provided for one’s hard work being enough for copyright protection to accrue. “‘Sweat of the brow’ courts thereby eschewed the most fundamental axiom of copyright law—that no one may copyright facts or ideas.” The Feist Court concluded: “In summary, the 1976 revisions to the Copyright Act leave no doubt that originality, not ‘sweat of the brow,’ is the touchstone of copyright protection in directories and other fact-based works. Nor is there any doubt that the same was true under the 1909 Act.”
In the case of the “Théâtre D’opéra Spatial” image discussed earlier, Allen claimed he tried at least 624 text prompts before arriving at the image he eventually entered into the art contest. Although the Copyright Review Board recognized that the prompts were not the point of creativity, they did not specifically call out the fact that Allen was essentially advancing a “sweat of the brow” argument:
It is the Office’s understanding that, because Midjourney does not treat text prompts as direct instructions, users may need to attempt hundreds of iterations before landing upon an image they find satisfactory. This appears to be the case for Mr. Allen, who experimented with over 600 prompts before he “select[ed] and crop[ped] out one ‘acceptable’ panel out of four potential images . . . (after hundreds were previously generated).” As the Office described in its March guidance, “when an AI technology receives solely a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology—not the human user.”
Simple text prompts as instructions are not the point of creativity because of longstanding precedent of copyright vesting with the artist and not the commissioner, and because of hard work and trial and error being inapplicable under the retired “sweat of the brow” doctrine. Accordingly, the Copyright Review Board and the courts are highly unlikely to recognize copyright in future AI-generated images solely on the prompts the user input into the machine. Uniquely original human contribution to the image may be protectible, but the image as a whole will not be entitled to protection by virtue of text prompts without anything more.
AI-generated images have raised existential aesthetic, societal, and moral questions about whether creativity is a uniquely human trait; what is the meaning of “art” and can a machine create it; will AI replace creatives; and, as touched upon in the Copyright Office’s “Guidance: Works Containing Material Generated by Artificial Intelligence,” should the use of AI be disclosed to others—and there are far more copyright questions that need answering as well. For accrual of copyright protection, though, it is settled law that human creativity is the “bedrock” and the “sine qua non” of the originality requirement, and without anything more, text prompts alone do not constitute the point of creativity upon which copyright attaches. In the eyes of the Copyright Act, creativity is a uniquely human quality.
