Copyright and AI: Key Takeaways from the U.S. Copyright Office’s Latest Report
BREAKING UPDATE
The intersection of artificial intelligence (AI) and copyright law is one of the most pressing legal questions industry leaders and their legal representatives face today. As generative AI tools become more sophisticated—creating everything from text and images to music and deepfake videos—there is increasing uncertainty about who (if anyone) owns the rights to AI-generated content.
In response to these challenges, the U.S. Copyright Office recently released Copyright and Artificial Intelligence: Part 2 – Copyrightability, a comprehensive report analyzing how existing copyright laws apply to AI-generated works and what, if any, legal changes may be needed. This report follows the first part of the study, which addressed digital replicas and AI-generated likenesses.
Below I break down the key findings of the Copyright Office’s latest report, analyze its implications for businesses and legal professionals, and explore the broader policy debate on AI-generated content.
Want to learn more? Register for the Web3 Ready Lawyer™ live web training.
1. The Fundamental Principle: Copyright Requires Human Authorship
The most significant conclusion from the report is a reaffirmation of the longstanding legal principle that copyright protection requires human authorship. This means that works created entirely by AI, without human creative input, are not copyrightable under current U.S. law.
This principle was reinforced by the 2023 case Thaler v. Perlmutter, where a federal court upheld the Copyright Office’s refusal to register an image created solely by an AI system, stating that “copyright law protects only works of human creation.” The new report makes clear that machines, no matter how advanced, cannot be “authors” under U.S. copyright law.
2. AI as a Tool: The Role of Human Contribution
While fully autonomous AI-generated works are not eligible for copyright, the report clarifies that AI-assisted works can still be protected if they involve sufficient human creativity.
The Copyright Office lays out three main ways that human authors can contribute to AI-generated works in a way that could warrant copyright protection:
Expressive Inputs – If a human provides original creative content, such as a detailed sketch or an unfinished manuscript that an AI system then modifies, that human input may be protected.
Modifications and Editing – A human who significantly revises or curates an AI-generated work to add creative expression may claim copyright protection over those modifications.
Selection and Arrangement – If a human selects, organizes, or arranges AI-generated content in a way that reflects originality and creativity, they may be able to claim copyright over that arrangement (but not the AI-generated elements themselves).
3. Prompts Alone Do Not Create Copyrightable Works
One of the most debated aspects of AI copyrightability is whether the person who writes a prompt for an AI system can claim copyright over the AI-generated output.
The report takes a firm stance: No, prompts alone are not enough to establish authorship.
Why? Because entering a prompt into an AI system is more like giving instructions than engaging in actual creative expression. The AI system is responsible for translating the input into an output, and users lack control over how the AI makes expressive decisions.
Even highly detailed prompts do not meet the threshold for authorship under U.S. copyright law because:
Prompts are instructions, not expressions – Giving directions to an AI is not the same as creating the work itself.
Outputs are unpredictable – The same prompt entered multiple times can generate vastly different outputs, demonstrating a lack of human control over the final result.
The “black box” problem – Many generative AI systems process prompts in ways that even their developers don’t fully understand, making it difficult to argue that the user meaningfully controls the creative process.
This means that AI-generated content is fundamentally different from human-created art, even when a human provides detailed instructions.
4. Copyright Office Rejects Special Protections for AI-Generated Content
The Copyright Office also weighed in on a major policy debate: should AI-generated works receive some form of sui generis (special, non-copyright) protection?
For now, the answer is no. The report found that existing copyright law is sufficient to address AI-related issues and that there is no compelling case for a new legal framework at this time.
Some stakeholders argue that AI-generated works should have their own system of protection, similar to how semiconductor chip designs and databases receive limited IP protections. However, the Copyright Office believes that such a change would require Congressional action and should not be rushed without further study.
5. Global Comparisons: How Other Countries Are Approaching AI Copyright
The U.S. Copyright Office’s position contrasts with some international approaches:
European Union – The EU’s AI Act is one of the most comprehensive regulatory efforts targeting AI, but it still follows the principle that copyright requires human authorship.
United Kingdom – The UK has a unique “computer-generated works” doctrine that allows AI-generated content to receive copyright protection for 50 years, but this approach has been criticized for being outdated.
China – China has taken an aggressive approach by recognizing AI-generated works as potentially protectable, though enforcement remains uncertain.
As the AI industry grows, global copyright policies may continue to evolve, and the U.S. will likely face pressure to clarify its legal stance further.
Key Takeaways for Businesses and Legal Professionals
For companies, content creators, and legal professionals working at the intersection of AI and intellectual property, this report provides crucial guidance:
· AI-assisted works can still be protected if there is sufficient human creativity.
· Users of generative AI tools cannot claim copyright simply by entering prompts.
· Fully autonomous AI-generated works are not eligible for copyright under current U.S. law.
· No new AI-specific copyright protections are planned for now.
· Businesses should consider contractual protections for AI-generated works (e.g., licensing agreements, trade secrets, or branding protections).
The Future of AI and Copyright
While the Copyright Office’s latest report provides clarity on current law, it also leaves many unanswered questions:
· Will Congress eventually introduce new legal protections for AI-generated works?
· How will courts handle disputes over AI-assisted authorship?
· Will companies push for more aggressive AI copyright reforms?
For now, businesses and legal professionals should focus on compliance with existing copyright laws while preparing for future regulatory shifts. AI is reshaping the creative landscape, and the legal system will have to keep evolving to address its impact.
Want to learn more? Register for the Web3 Ready Lawyer™ live web training.
· Attract & retain high-value Web3 clients
· Spot legal issues before they become malpractice risks
· Position yourself as a top-tier advisor in the Web3 economy
Thanks for breaking that down for us. As someone who uses AI on regular basis, it’s good to know what’s what.
I was wondering though, what’s your take on AI using data available but not always crediting the output? I feel there’s still a debate out there about using AI’s generated content when you don’t know the source.