Content generated solely by artificial intelligence is not protected by copyright law, according to new guidelines issued by the US Copyright Office.
The announcement was included in a sweeping report on policy issues related to artificial intelligence, particularly its ability to be copyrighted—a flashpoint within creative industries in recent years.
The document states that despite its ongoing advancements, generative AI is liable to the United States’s current copyright principles, which take a strict stance on what material—human- or machine-made—qualifies for protection.
Per the guidelines, AI prompts, such as text-to-image generator tools, don’t currently permit sufficient control to “make users of an AI system the authors of the output.” This applies unilaterally to simple text prompts and complex instruction. “No matter how many times a prompt is revised and resubmitted, the final output reflects the user’s acceptance of the AI system’s interpretation, rather than authorship of the expression it contains,” the report states.
The report illustrated how easily matters of authorship can be muddled with AI using the example of a Gemini-generated cat smoking a pipe and reading a newspaper. Gemini, the reported noted, ignored select prompt instructions and added elements of its own, including the cat’s “incongruous human hand.”
The unpredictability of Gemini was then contrasted with examples of human spontaneity, like the splatter technique of Jackson Pollock. The artist did not manage where or how the paint landed, but “controlled the choice of colors, number of layers, depth of texture, placement of each addition to the overall composition — and used his own body movements to execute each of these choices.”
“The issue is the degree of human control, rather than the predictability of the outcome,” the office concluded.
However, the department said that using such technology to assist in “human” creative expression does not necessarily preclude a work’s eligibility for copyright protection. Like a writer asking AI to create an outline for a book, the user is “referencing, but not incorporating, the output,” the report explained.
In 2022, for example, author Kris Kashtanova claimed to have received a copyright for an AI-created work when her request to register her comic book Zarya of the Dawn was approved. The Copyright Office, however, put its decision under review and requested additional information when it was discovered that the book’s images had been made using AI generator Midjourney.
The Copyright Office ultimately cancelled its original certification and issued a new one solely covering the elements that Kashtanova created.
The principle also applies to artists who use AI systems to modify preexisting work, like tweaking characters, or adding subtle design elements to an illustration. The AI-generated elements would be excluded from the copyright, but if the human-made product remains recognizable, the “perceptible human expression” can still be copyrighted.
Works that incorporate AI-generated elements can be protected if there’s been perceptible creative modification—if, say, an artist significantly rearranges elements of an AI-artwork or pairs it with text written by a human. In the same spirit, “a film that includes AI-generated special effects or background artwork is copyrightable, even if the AI effects and artwork separately are not.”
The Copyright Office added that prompt-generated images could even receive protection—stressing the “case-by-case determination”— if the user adequately modified parts of the picture. In this case, the AI work could be considered like any derivative artwork, just derivative of a machine, rather than a human.
Since platforms such OpenAI Inc.’s ChatGPT, DALL-E, and Midjourney became accessible to the public, requests to copyright works with AI have risen dramatically—and so have legal challenges from digital artists who claim the platforms plagiarize human-made work. The Copyright Office has seemingly found a legal precedent in the issue of appropriation art, a tradition in which one artist repurposes another’s creation.
Rulings in these cases often hinge on whether modifications to the original work have been deemed as sufficiently “transformative.” AI has raised those stakes. Where it was once two artists in contention, courts must now factor in the diffusion of millions of digital artworks by generative platforms, and, given the new guidelines, determine the boundaries “human expression.”
But even these principles, the report concludes, could change if AI evolves beyond past legal precedents: “In theory, AI systems could someday allow users to exert so much control over how their expression is reflected in an output that the system’s contribution would become rote or mechanical.”
