The U.S. Copyright Office (Copyright Office) has released Part 1 and Part 2 of its 3-Part Report on Copyright and Artificial Intelligence (Report), analyzing the legal and policy challenges surrounding artificial intelligence (AI), Digital Replicas and Copyrightability. Parts 1 and 2 of the Report affirm existing copyright principles while also calling for new federal protections against unauthorized AI-generated likenesses.
The Copyright Office is preparing to release Part 3, which will focus on the legal implications of training AI models on copyrighted works, including licensing considerations and liability allocation.
Additionally, the Office announced plans to update its March 2023 registration guidance regarding works containing AI-generated content and revise sections of the Compendium of U.S. Copyright Office Practices to address AI-generated content based on the conclusions of Part 2 of the Report.
Key Takeaways from the Report
Digital Replicas – The Case for New Federal Protections (Part 1) The proliferation of AI-generated deepfakes and synthetic media has raised significant concerns about privacy, consent, and misrepresentation. However, as the Copyright Office “copyright law protects only original expression, not facts or ideas—including the identity or likeness of an individual.” As a result, copyright does not provide a remedy for individuals whose AI-generated likeness is used without authorization.
Recognizing this critical gap in legal protections, the Office urges Congress to enact a new federal law to protect individuals against the unauthorized distribution of digital replicas. The Copyright Office proposes that Congress enact law that would:
- Apply to all individuals, not just celebrities or public figures.
- Prohibit the known distribution of unauthorized digital replicas, even when the use is non-commercial.
- Mandate that online platforms remove unauthorized digital replicas upon notification, creating a safe harbor framework similar to the copyright takedown systems.
- Provide robust enforcement mechanisms, including injunctive relief, statutory damages, and potential criminal penalties for severe violations.
The Office highlights the urgency of this issue, emphasizing the growing concern over digital replicas as AI technology advances. Without federal protection, victims of AI-generated impersonation face legal uncertainty when seeking recourse.
Copyrightability of AI-Generated Content (Part 2)
Part 2 of the Report focuses on whether AI-generated works qualify for copyright protection and reaffirms that human authorship required under existing copyright law remains essential. The Copyright Office distinguishes between using AI as an assistive tool in the creation of works and as a replacement for human creativity, making it clear that, “AI-generated material cannot be copyrighted unless a human author has determined the expressive elements of the work.”
Key clarifications in Part 2 include:
- Text prompts alone do not establish authorship. The Copyright Office explains that simply entering prompts into a generative AI system does not provide the human authorship needed for copyright protection.
- Copyrightability is analyzed on a case-by-case basis. AI-human hybrid works may qualify for protection if a human modifies, arranges, or exercises creative control over AI-generated content. If a human inputs their existing copyrightable work into a generative AI system (beyond just text prompts), they may be considered the author – at least to the extent the original work remains perceptible in the output.
- AI-assisted creation is allowed but fully AI-generated works cannot. AI may be used as a creative tool similar to photography or digital editing software, but the final work must reflect human authorship to qualify for copyright protection
The Copyright Office rejects proposals to introduce new copyright protections for purely AI-generated content, emphasizing that, “copyright law has long adapted to new technologies—from photography to software—and there is no need to craft a separate framework.” As a result, AI-generated content remains ineligible for copyright protection unless it incorporates a human’s creative contributions.
What’s Next? The Copyright Office will release Part 3 of the Report, but no publication date has been set. Part 3 is expected to address the use of copyrighted works in AI model training datasets, which is one of the most contentious issues in AI and copyright. It is also expected to:
- Examine legal liability for AI developers that train models on copyrighted content.
- Explore licensing frameworks for copyrighted material used in AI model training and whether fair use applies.
- Analyze potential regulatory responses to AI’s reliance on copyrighted material.
In addition, based on the findings of the Report, the Copyright Office plans to update its March 2023 registration guidance regarding works containing AI-generated content. It also plans to revise the “Compendium of U.S. Copyright Office Practices” to clarify how AI-assisted works are registered and evaluated under copyright law.
Implications for AI Developers, Content Creators Using AI, and Online Platforms
The Copyright Office’s Report highlights major legal and regulatory considerations for companies creating, distributing, or using AI-generated content, as well as those developing generative AI systems:
- AI-generated content is not automatically copyrightable. Businesses should ensure that human authorship is evident in workflows that use AI if they intend to seek copyright protection.
- Changes to the copyrightability of AI-generated material may impact agreements with third parties. Company copyright strategy and contracts should be aligned with the Copyright Office’s current positions on the copyrightability of AI-generated content.
- Federal digital replica laws could impact AI companies and online platforms. If Congress enacts the Copyright Office’s recommendations, companies developing deepfake, voice synthesis, or digital identity tools may face new compliance obligations and legal risks. Further, content hosts and social media services could be legally obligated to remove unauthorized digital replicas upon request.
Next Steps & How We Can Help
With AI regulation evolving rapidly, companies must proactively assess their compliance risks and copyright strategies. We recommend:
- Reviewing AI-generated content policies. Ensure that AI-generated material is created and used in adherence to current copyright principles and does not rely on copyright protection where none applies.
- Preparing for federal regulation of digital replicas. Online platforms using synthetic likenesses and voice replication should assess potential risks and implement compliance strategies.
- Keeping informed about Copyright Office Actions. Watch for the release of Part 3 of the Copyright Office’s Report covering training AI models on copyrighted works and for upcoming updates to the Compendium of U.S. Copyright Office Practices because it may impact the registration process for AI-assisted works and fair use considerations.
The CommLaw Group Can Help!
With sweeping changes in U.S. AI policy, emerging state laws, and the rapid evolution of AI technologies, businesses need experienced counsel to navigate the challenges ahead. The CommLaw Group offers deep expertise in AI and privacy compliance, helping companies adapt to new regulations, mitigate risks, and maintain a competitive edge.
CONTACT US TODAY
Our team is ready to guide your company through the complexities of AI and privacy governance.
BrianAlexander: 703-791-1050 | bal@commlawgroup.com
Susan Duarte: 703-714-1318 | sfd@commlawgroup.com
Diana James: 703-663-6757 | daj@commlawgroup.com
A special thanks to our law clerk, Fusheng Zhou, who is a law student at George Washington School of Law, for his work in preparing this advisory.
