Copyright Report Says AI-Generated Songs Aren’t Protected, But That AI Tools Can Still Be Fair Game
Written by djfrosty on January 29, 2025
A new federal report on artificial intelligence says that merely prompting a computer to write a song isn’t enough to secure a copyright on the resulting track — but that using AI as a “brainstorming tool” or to assist in a recording studio would be fair game.
In a long-awaited report issued Wednesday (Jan. 29), the U.S. Copyright Office reiterated the agency’s basic stance on legal protections for AI-generated works: That only human authors are eligible for copyrights, but that material created with the assistance of AI can qualify on a case-by-case basis.
Amid the surging growth of AI technology over the past two years, the question of copyright coverage for outputs has loomed large for the nascent industry, since works that aren’t protected by copyrights would be far harder for their creators to monetize.
Trending on Billboard
“Where that [human] creativity is expressed through the use of AI systems, it continues to enjoy protection,” said Shira Perlmutter, Register of Copyrights, in the report. “Extending protection to material whose expressive elements are determined by a machine, however, would undermine rather than further the constitutional goals of copyright.”
Simply using a written prompt to order an AI model to spit out an entire song or other work would fail that test, the Copyright Office said. The report directly quoted from a comment submitted by Universal Music Group, which likened that scenario to “someone who tells a musician friend to ‘write me a pretty love song in a major key’ and then falsely claims co-ownership.”
“Prompts alone do not provide sufficient human control to make users of an AI system the authors of the output,” the agency wrote. “Prompts essentially function as instructions that convey unprotectible ideas.”
But the agency also made clear that using AI to help create new works would not automatically void copyright protection — and that when AI “functions as an assistive tool” that helps a person express themselves, the final output would “in many circumstances” still be protected.
“There is an important distinction between using AI as a tool to assist in the creation of works and using AI as a stand-in for human creativity,” the Office wrote.
To make that point, the report cited specific examples that would likely be fair game, including Hollywood studios using AI-powered tech to “de-age” actors in movies. The report also said AI could be used as a “brainstorming tool,” quoting from a Recording Academy submission that said artists are currently using AI to “assist them in creating new music.”
“In these cases, the user appears to be prompting a generative AI system and referencing, but not incorporating, the output in the development of her own work of authorship,” the agency wrote. “Using AI in this way should not affect the copyrightability of the resulting human-authored work.”
Wednesday’s report, like previous statements from the Copyright Office on AI, offered broad guidance but avoided hard-and-fast rules. Songs and other works that use AI will require “case-by-case determinations,” the agency said, as to whether they “reflect sufficient human contribution” to merit copyright protection. The exact legal framework for deciding such cases was not laid out in the report.
The new study on copyrightability is the second of three studies the agency is conducting on AI. The first report, issued last year, recommended federal legislation banning the use of AI to create fake replicas of real people; bills that would do so are pending before Congress.
The final report, set for release at some point in the future, deals with the biggest AI legal question of all: whether AI companies break the law when they “train” their models on vast quantities of copyrighted works. That question — which could implicate trillions of dollars in damages and exert a profound effect on future AI development — is already the subject of widespread litigation.