AI and Copyright: Key Takeaways from Part 2 of the U.S. Copyright Office AI Report

Artificial intelligence (AI) is transforming how we create, from generating art and music to drafting articles and code. But who owns the rights to AI-generated content? The U.S. Copyright Office recently released a landmark report, Copyright and Artificial Intelligence, Part 2: Copyrightability, to answer this critical question. This article breaks down the key takeaways of the report, and its implications for content creators, businesses, AI developers, and the broader legal landscape.

Key Takeaways from the Report

1. Human Authorship is Essential

Copyright law exists to incentivize human creativity. AI cannot hold rights because it lacks consciousness and does not require financial or reputational incentives to create. Granting copyright to AI-generated works would dilute the law’s core purpose: fostering human creativity. If a work lacks meaningful human authorship (e.g., an image produced solely by typing “create a sunset photo” into an AI tool), it won’t qualify for protection. The report cites Thaler v. Perlmutter, where a work created solely by AI was denied copyright, reinforcing that non-human entities cannot hold copyright protection.

2. AI is a Tool, Not a Creator

The report distinguishes between AI as a tool for human creators and AI as an autonomous creator. When AI is used merely to assist in the creative process, such as enhancing existing content or executing specific artistic choices, the final output may still qualify for copyright protection. However, the work is not copyrightable if AI operates independently, generating content with little to no human direction.

3. The Rejection of “Sweat of the Brow” Arguments

The “sweat of the brow” doctrine (rewarding labor over creativity) was rejected in Feist v. Rural Telephone. Likewise, simply spending hours refining AI prompts does not establish copyright. Only original human expression qualifies. This principle is crucial for AI-generated works where human involvement is limited to iterative prompting without genuine creative contribution.

4. Prompts Alone Are Insufficient for Copyright Protection

Prompts alone do not constitute copyrightable authorship. Simply typing prompts (e.g., “Write a play in the style of Wole Soyinka set in 1950s Nigeria”) doesn’t make you the author of the AI’s output. Prompts are instructions conveying ideas that are not protectable under copyright law. Additionally, the unpredictable nature of AI outputs, where identical prompts can generate different results, weakens any claim that prompting alone demonstrates authorship.

When Does AI-Assisted Work Qualify for Copyright?

The Copyright Office uses a case-by-case analysis, focusing on two criteria:

1. Human Contribution Rule

For AI-assisted work to qualify for copyright, a human must exercise creative control over the work’s expression. This requires making meaningful artistic decisions rather than merely selecting or refining AI-generated content. For example, if an artist generates base images using AI but manually edits the composition, colour, and lighting to align with their artistic vision, the work may qualify for protection. In contrast, if a user creates a logo using AI and only adjusts the font size, their involvement is deemed insufficient for copyright protection.

2. Identifiable Human Contribution

Copyright applies only to portions of a work that can be directly attributed to human creativity. The Copyright Office assesses whether a human’s input is substantial enough to warrant protection. A notable case illustrating this principle is Zarya of the Dawn (2023), where the U.S. Copyright Office revoked copyright protection for AI-generated images in a comic book. Although the text was human-authored and remained protected, the images, created using Midjourney, were considered ineligible for copyright. This decision reinforces the stance that AI-generated content, absent significant human intervention, does not meet the legal requirements for authorship.

Why the Copyright Office’s Rejects Legal Reform

The U.S. Copyright Office believes that existing copyright laws are enough to handle AI-generated works, so there is no need to change the law. Their main argument is that copyright exists to protect and encourage human creativity, not machine-generated content. Since AI does not need financial incentives, granting copyright protection to AI-generated works could weaken the law’s original goal of promoting human authorship.

The Copyright Office also rejects the idea of creating a special legal system just for AI-generated works, arguing that it would make copyright law more complicated without any real benefit. Another major concern is that allowing AI-generated works to be copyrighted could flood the market with AI content, reducing the value of works created by humans.

Additionally, the Copyright Office finds no compelling evidence that excluding AI-generated works from copyright protection puts the United States at a competitive disadvantage in the global AI race. Given these factors, it concludes that current copyright laws remain well-equipped to address the challenges posed by AI-generated content, making legal reform unnecessary at this time.

What This Means for You

1. For Content Creators

• Human Authorship is Essential: Copyright protection applies only to works with significant human authorship. If AI-generated content is merely prompted without meaningful human input, it does not qualify for protection. Creators must actively contribute to the final work beyond just using an AI tool.
• AI is an Assistive Tool, Not a Replacement: Creators can still use AI to assist in the creative process, but the AI's role must be genuinely assistive, not a replacement for human creativity. Examples include using AI for brainstorming, enhancing existing work, or modifying specific aspects of a work. Simply generating content through AI and claiming it as your own does not meet copyright standards.
• Expressive Inputs Can Qualify for Copyright: If creators include their own creative works, such as drawings, photos, or music, as inputs to an AI system, and those elements are discernible in the output, they are copyrightable. This is similar to a derivative work, where the copyright extends to the new human elements, not the AI-generated aspects
• Creative Modifications Make AI Outputs Protectable: Copyright may apply if a creator significantly transforms AI-generated content. This could involve editing and refining AI-generated text or images and selecting and arranging AI-generated elements uniquely, adapting AI outputs to create a cohesive, human-directed final work. If a human author modifies or arranges AI-generated content creatively, the selection, coordination, and arrangement of the elements can be eligible for copyright protection, so long as it meets the standard for originality
• Documentation Required: When registering works with the Copyright Office, creators must disclose any use of AI in the creative process, and specify their human contributions, such as editing, arranging, or modifying AI outputs. Failing to disclose AI involvement could lead to legal complications later.
• Ethical Considerations: Be transparent about AI use. Audiences and clients increasingly care about authenticity, and misrepresenting AI-generated content as human-made could damage trust.

2. AI Developers and Tech Companies

• Incentives Exist: The lack of copyright protection for AI-generated outputs, such as text, images, or music created by AI, does not pose a major barrier to innovation. Developers can rely on patents to protect innovative processes or inventions, trade secrets to safeguard confidential business information, and copyrights to protect the code and software systems they create. These existing forms of intellectual property (IP) protection ensure that developers can still benefit from their efforts and investments.
• Focus on Tools That Assist Humans: AI developers can succeed by creating tools that enhance human creativity rather than replace it. They can partner with artists, writers, and musicians to build tools that help them create better or faster. They should also develop AI systems that modify or improve creative works, making them more appealing to creators. By focusing on assistive tools, developers can ensure their products are widely adopted and valued by creative professionals.
• Transparency is Key: To build trust and reduce legal risks, AI developers should clearly explain how their AI models are trained (e.g., what data is used). They should also be open about the capabilities and limitations of their systems.
• Potential Liability: AI developers may face liability issues if their AI systems infringe on existing copyrights. The Report acknowledges that a subsequent part will address the training of AI models on copyrighted works, licensing considerations, and allocation of any liability.

3. For Businesses

• Clear Understanding of Copyrightability: Businesses must understand that copyright protection is not automatic for AI-generated content. Companies relying heavily on AI-generated content for marketing or product development need to revise their workflows to include substantial human creativity.
• Focus on Human-Centered Design: Businesses should prioritize workflows that involve human authors in the AI process. This could involve incorporating expressive inputs, human modifications, or creative arrangements of the AI-generated content. For example, a company using AI to draft blog posts should have editors significantly rewrite outputs to claim copyright.
• Due Diligence: Businesses should ensure that any use of AI-generated content does not infringe upon existing copyrights and that they are not passing off non-protected content as original works. AI training data often includes copyrighted material. Ensure your AI tools use properly licensed datasets to avoid infringement claims.
• Monitoring International Standards: Businesses should also track how the international landscape addresses these issues. As countries take different approaches, this will matter for companies engaged in international business.

4. Legal Professionals

• New Litigation Areas: Courts are expected to see an increase in cases that challenge the boundaries of “human authorship” in works created with the help of AI. A key example is the Zarya of the Dawn ruling. This case highlights the importance of understanding how courts and the Copyright Office view AI-assisted works. Lawyers should stay informed about such precedents to navigate this evolving legal landscape effectively.
• Focus on Human Control: A critical factor in these cases will be the degree of human control over the creative process. Lawyers will need to argue how much human input shaped the expressive elements of AI-generated outputs. However, this can be challenging, as many current AI tools operate in ways that make it difficult to clearly define or quantify human involvement. Establishing a strong connection between human creativity and the final work will be essential for securing copyright protection.
• Guidance on Registration: The Copyright Office is expected to release new guidance on how to register works involving AI, which will be a critical resource for lawyers advising clients. This guidance will help clarify what level of human involvement is necessary for copyright eligibility, ensuring that works with sufficient human authorship are properly protected.
• Monitoring Policy Changes: As AI technology continues to evolve, so too will the legal and policy landscape surrounding it. Lawyers must stay vigilant and monitor any changes in regulations, court rulings, or Copyright Office policies. Keeping up with these developments will be key to providing accurate advice and helping clients adapt to new challenges in this rapidly changing field.

Conclusion

The U.S. Copyright Office’s stance on AI-generated content reaffirms a fundamental principle of copyright law: human creativity remains at the core of protectable authorship. While revolutionary, AI is seen as a tool to enhance human efforts rather than an independent creator deserving of copyright protection.

As AI continues to revolutionise creative industries, stakeholders must adapt their strategies to align with existing copyright frameworks. For content creators, this means adapting workflows to emphasize human creativity, for AI developers, it involves focusing on tools that complement human ingenuity, for businesses, a careful approach to intellectual property management, and for legal professionals, a readiness to navigate new legal territories. Legal professionals, meanwhile, must stay ahead of emerging case law and policy changes to provide sound guidance in this rapidly evolving field.

However, as AI capabilities continue to advance, future legal and policy debates will likely revisit the balance between human authorship and machine-generated content. The dialogue surrounding AI and copyright is far from settled, making it essential for all stakeholders to remain engaged with ongoing developments.

*This blog post is for informational purposes only and does not constitute legal advice.

*This blog post was written with the aid of generative AI tools.