Public Invited to Provide Input on AI and Content Ownership to US Copyright Office
The US Copyright Office (USCO) is seeking public input on generative AI and the ownership of its outputs. This technology has become a focal point in the legal system, prompting the office to solicit public comments on various complex issues. These include concerns about companies using copyrighted material to train AI models, determining copyright eligibility for AI-generated content, addressing liability for copyright infringement, and managing machine-generated outputs that imitate the work of human artists.
“The adoption and use of generative AI systems by millions of Americans — and the resulting volume of AI-produced material — has sparked widespread public debate about what these systems could mean for the future of the creative industries and raises significant questions for the copyright system,” the USCO wrote in a statement released Wednesday.
One issue the office hopes to address is requiring human authorship to register copyright in (otherwise AI-based) content, drawing increasing efforts to protect copyrighted material that names an AI as the author or author. “The crucial question seems to be whether the ‘work’ is essentially a taking of human authorship, with the computer merely an aid, or whether the traditional elements of authorship (literary, artistic or musical expression or elements of choice, arrangement, etc.) of the work are in fact not designed by man but by machine and implemented,” USCO wrote.
While the issue is far from settled, several cases have provided clues as to where the limits may fall. For example, the agency said in February that the text and layout of a (human) graphic novel created in part by artificial intelligence is copyrighted, but the images created by Midjourney in the work are not. On the other hand, a federal judge recently rejected an attempt to register AI-generated art with no human intervention other than a provocative text prompt. “Copyright has never stretched so far […] to protect works created by new forms of technology that operate without a guiding human hand, as plaintiff here urges,” U.S. District Judge Beryl Howell wrote in the ruling.
The USCO is also seeking input into the increasing number of infringement lawsuits by copyright owners against AI companies for training their published works. Sarah Silverman is one of the high-profile plaintiffs suing OpenAI and Meta for allegedly training ChatGPT and LLaMA (respectively) in their written work—in her case, her 2010 memoir The Bedwetter. OpenAI is also facing a class action lawsuit over using copied web data to train a viral chatbot.
The USPO says the public comment period is open until November 15. Until then, feel free to share your thoughts.
