AI and Deepfakes: U.S. Copyright Office Urges Federal Digital Replica Law
In Short
The Situation: In early 2023, the U.S. Copyright Office ("Office") launched a new initiative to examine the intersection of copyright law and artificial intelligence ("AI"). Later that year, the Office issued Registration Guidance on Works Containing Material Generated by Artificial Intelligence and published a Notice of Inquiry ("NOI") seeking public comment on a number of issues related to copyright and AI.
The Development: On July 31, 2024, the U.S. Copyright Office published Part One of its report on AI and copyright, which focuses on unauthorized digital replicas or "deepfakes"—i.e., simulations that use AI to replicate a person's appearance and/or voice. Part One concludes that there is an "urgent need" for a federal law addressing unauthorized digital replicas.
Looking Ahead: Stakeholders should watch for federal and state proposals to address digital replicas, as well as other legislative activity related to AI and copyright. In addition, the U.S. Copyright Office has advised that forthcoming Parts of its report are expected to cover other issues raised in the NOI, including the copyrightability of works created using generative AI, training of AI models on copyrighted works, licensing considerations, and allocation of any potential liability. These Parts will be published as they are completed.
The U.S. Copyright Office ("Office") recently published Part One of its Copyright and Artificial Intelligence Report ("Part One"), recommending a new federal law addressing digital replicas or deepfakes made to mimic an individual's image, voice, or likeness—which have become easier to make with the fast-growing and widespread availability of generative AI ("GenAI") tools. Shira Perlmutter, Register of Copyrights and Director of the U.S. Copyright Office noted that digital replicas pose a "serious threat" to not just entertainers and politicians but also private citizens, and concluded that there is "an urgent need" for federal legislation to address serious potential harms to "reputations and livelihoods."
In Part One, the Office surveyed various federal and state efforts to regulate AI and the risk of digital replicas and concluded that existing laws are inconsistent and insufficient to address the harms posed by digital replicas in the age of AI. The Office found that existing federal laws (including the Copyright Act, Federal Trade Commission Act, Lanham Act, and Federal Communications Commission Act) are "too narrowly drawn to fully address the harm from today's sophisticated digital replicas." And while some states have passed laws addressing digital replicas (including Tennessee, New York, and Louisiana), there is variation among these laws as well as other state laws that may have arguable applicability in this context (e.g., laws addressing rights to publicity or privacy).
Against this backdrop, the Office recommended that Congress establish a federal law that protects all individuals during their lifetimes from the knowing distribution of unauthorized digital replicas. The Office also advised that Congress incorporate these concepts into that legislation:
- Comprehensively define digital replica to include video, image, and audio recordings that have been digitally created or manipulated to realistically but falsely depict an individual;
- Protect all persons—not just celebrities or public figures;
- Guarantee protection for at least the individual's lifetime;
- Introduce liability for distributing or making available an unauthorized digital replica;
- Create secondary liability for online service providers who knowingly and materially contribute to infringing conduct, combined with an appropriate safe harbor;
- Permit licensing of digital replica rights, but not their assignment;
- Address First Amendment concerns through the use of a multifactor balancing test;
- Provide injunctive relief and monetary damages; and
- Advise against full preemption of state laws.
The release of Part One represents another significant step taken by the Office to address the complexities and uncertainty at the intersection of copyright law and AI, after receiving more than 10,000 comments from a broad range of interested parties (across industries, all 50 U.S. states, and 67 countries) in response to the Office's NOI last year. Additional activity from the Office is expected this year, including the Office's release of additional Parts of its Copyright and Artificial Intelligence Report.
Three Key Takeaways
- Part One notes that current federal law, including the U.S. Copyright Act, does not provide specific protection against unauthorized digital replicas. Although there are several federal and state laws and common-law doctrines that provide a patchwork of protection, none sufficiently address the sweeping implications of unauthorized digital replicas.
- The Office recommends that Congress establish a federal law that protects all individuals during their lifetimes from the knowing distribution of unauthorized digital replicas. Liability should arise from the distribution or making available of an unauthorized digital replica, but not the act of creation alone. The Office advises that new federal legislation should not be limited solely to commercial uses, as the harms caused by digital replicas are often personal in nature.
- The Office will release subsequent Parts of its report on AI and copyright. Forthcoming report topics include the copyrightability of works created using GenAI, training of AI models on copyrighted works, licensing considerations, and allocation of any potential liability.
