Copyright Office Faces "Technical Debt" Over AI-Generated Works
The U.S. Copyright Office's decades-long refusal to define when human-machine collaboration merits copyright protection has created a legal quagmire for AI-generated works, according to Harvard Law School Professor Rebecca Tushnet.
The office accepted copyright claims for machine-assisted creations-security camera footage, real estate photographs-without requiring substantial human involvement, Tushnet said during a panel discussion at Harvard Law School. That accumulated indecision now complicates decisions about whether AI-generated images, text, and other outputs deserve legal protection.
"Because we decided to stop thinking about what level of human-machine interaction leaves the human in charge enough to deserve copyright, that's why we're struggling now," Tushnet said.
The Supreme Court Reinforces Human Authorship
The Supreme Court declined in March to hear Thaler v. Perlmutter, letting stand lower court rulings that upheld the Copyright Office's position: copyright law requires human authorship. The decision closes one avenue for challenging the office's stance on AI outputs.
The Copyright Office has since required applicants to disclose any use of generative AI and LLM in their work. Jessica Silbey, Boston University School of Law professor, predicted the office will demand greater detail about AI use in future applications, once staffing levels improve.
Mixed Results in Recent Cases
The Copyright Office has issued inconsistent rulings. A comic book writer registering "Zarya of the Dawn" could copyright the text and the "selection, coordination, and arrangement" of images, but not the AI-generated images themselves. The office denied copyright protection to the pictures because they were "not the product of human authorship."
In contrast, the office awarded copyright for "A Single Piece of American Cheese," a digital artwork where the human creator submitted evidence of 35 rounds of edits. Silbey called this "an amazing lawyering moment" and a signal of what the Copyright Office may prioritize going forward.
The Original Purpose: Incentivizing Human Creation
Copyright law, enacted in 1790, aimed to encourage human creativity by offering legal protection. Extending that protection to machine-generated works conflicts with the statute's foundational purpose, Silbey said.
Christopher Bavitz, Harvard Law professor and managing director of the Cyberlaw Clinic, argued that limiting AI copyright protection serves the broader goal of copyright law. "If I'm going to create things, it is helpful and beneficial for me if I have a corpus of things that I can draw upon like the great public domain works," he said. Balancing some protection for creators against a robust public domain remains central to copyright's function.
Industry Standards May Move Faster Than Law
Entertainment, music, and video game industries are already establishing their own norms around AI use, independent of copyright doctrine. In May, Hollywood studios and SAG-AFTRA, the labor union representing over 150,000 actors and media professionals, agreed to limits on AI-generated content in a four-year contract.
Tushnet described these industry agreements as "perfectly reasonable labor policy" driven by money and copyright concerns. Silbey added that practice-driven specificity is emerging across sectors. "The industries are going to start deciding before the law does what kind of AI we are going to say is creative and is acceptable and what is not," she said.
Bigger Questions Beyond Copyright
Copyright protection alone cannot address the deeper concerns around AI-generated content, Silbey cautioned. The real questions are philosophical: whether the world should contain material that mimics human thought without being produced by humans.
"Copyright is not going to solve that problem," she said. "We can either grant it all copyright or not grant it all copyright. It still may be produced and it's still mysterious."
For AI for Legal professionals, these cases signal that copyright doctrine will continue shifting as AI capabilities advance. Courts and regulators will likely require clearer disclosure of AI involvement and greater evidence of human creative control.
Your membership also unlocks:
