Register of Copyrights Shira Perlmutter testified in front of the Senate Judiciary Committee’s Subcommittee on Intellectual Property on November 13. Perlmutter told the Senators that the Copyright Office is still working to get parts two and three of its report on how AI impacts copyright by the end of 2024.
The first part of the report was released in July 2024 and focused mainly on the impact of digital replicas created by AI on current copyright laws. That report suggested a new federal law that would enshrine property rights for a person’s digital replica and with the aim of preventing false depictions of an individual created without consent.
That report resulted in the NO FAKES Act being introduced by the U.S. House—aiming to enshrine the right for individuals to control their digital replicas including voice and likeness. Perlmutter says the No FAKES Act is a step in the right direction, since it establishes a notice and takedown process for removing unauthorized replicas. Online platforms would also need to designate an agent as the contact point for reporting violations to keep from being held accountable for unauthorized reproductions of a person.
Subcommittee Chair Chris Coons (D-DE) asked Perlmutter “what keeps her up at night” when it comes to the AI issue, to which Perlmutter answered the Senator, “the speed at which this is all developing.” Perlmutter says she’s confident that copyright and AI issues will eventually be solved, but that she’s “less comfortable about what it means for humankind.”
House Administration Chairman Bryan Steil (R-WI) asked the Copyright Office to clarify when the rest of its AI report would be available. “The importance of these reports cannot be overstated,” Steil said in the letter. “The absence of these reports creates uncertainty for industries that are already grappling with AI-related challenges and hinders lawmakers’ ability to craft effective policy.”
In response at the hearing, Perlmutter says the U.S. Copyright Office has been “Trying to set and follow our own ambitious deadlines” with the goal remaining to get the rest of the report out by the end of 2024. Perlmutter says her chief concern is that the report be “accurate and thoughtful.”
The forthcoming report will include recommendations on dealing with the copyrightability of materials created using generative AI. It will also detail the legal implications of training generative AI on copyrighted works. The latter point is the most controversial—genAI companies like OpenAI and Anthropic have argued that their training material falls under the ‘fair use’ provision of current copyright law.
Perlmutter says the U.S. Copyright Office received more than 10,000 comments on its Notice of Inquiry issued in August 2023. The inquiry sought public comments on the intersection of copyright law and artificial intelligence.
Perlmutter testified that nearly all commenters said the existing fair use framework is adequate to address issues surrounding the use of copyrighted works in training generative AI models. She says the views diverge “strongly on what types of uses qualify as fair use,” however. Those issues surround the transparency of generative AI companies sharing materials used to train their models.