Copyright

As artificial intelligence (AI) technologies continue to evolve, copyright law is being challenged to keep pace. In the United States—a common law system where legal precedent plays a critical role—copyright protections around AI-generated content are still developing. The U.S. Copyright Office and courts have issued some guidance, but much remains unsettled.

This guide provides a foundational understanding of how copyright intersects with AI, including what is currently protected under U.S. law, examples of human-AI collaborations, and ongoing legal debates.

Under U.S. copyright law, protection is granted to original works that are independently created by a human author. This principle has been affirmed by court decisions, such as:

 

• Naruto v. Slater (2016)
  A macaque monkey took a selfie using a photographer’s camera. The court held that the monkey could not claim copyright, reinforcing the requirement of human authorship.

Key Takeaway:

AI-generated content that lacks human creative input is not eligible for copyright protection.

Human-AI Collaboration

Although works created solely by AI are generally not protected, there are cases where AI-assisted works have received partial copyright protection due to substantial human contribution.

Examples:

• A Single Piece of American Cheese (2024) by Kent Keirsey
  Created with the AI tool Invoke, this artwork involved over 35 human-guided AI edits using a process called inpainting. The arrangement and selection of elements qualified for copyright protection—even if the generated images themselves did not. Above, you can watch the time lapse creation of the work in Invoke.

 

 

• Zarya of the Dawn by Kristina Kashtanova
  A graphic novel with AI-generated illustrations (via Midjourney). The U.S. Copyright Office did not grant copyright to the AI images but did protect:
  • The original story text.
  • The compilation of text and images as a whole.

Important Concepts:

• Selection
• Coordination
• Arrangement

These elements can establish the required level of human creativity for copyright.

Inputs: Training Data

Training AI models—especially large language models (LLMs)—requires enormous datasets. These often include copyrighted content found online.

Legal Debate:
Some developers argue that this constitutes fair use, specifically transformative fair use.

Legal Precedents:

• Authors Guild v. HathiTrust (2014)
• Authors Guild v. Google (2015)
  In both cases, the courts upheld the digitization of copyrighted books for machine learning purposes as fair use. While these did not involve generative AI directly, they are relevant to the debate.

Outputs: AI-Generated Content

General Rule:

AI-generated outputs without meaningful human input are not copyrightable due to the lack of human authorship.

However, when a human significantly contributes to the creative process (e.g., through prompts, edits, curation), the final product may be eligible for protection.

Caution for Users:

• AI outputs are not guaranteed to be accurate.
• Some may include “hallucinations”—false or misleading information.
• Outputs may infringe on existing copyrights if substantially similar to protected works.

Legal Example:

• Concord Music Group, Inc. v. Anthropic PBC (2023)
  A complaint alleged that Anthropic’s AI model generated “identical or nearly identical” lyrics from songs like Gloria Gaynor’s “I Will Survive” and Katy Perry’s “Roar.”

AI is reshaping the creative landscape, but the legal system is still catching up. As we navigate the evolving rules around authorship, ownership, and fair use, it’s important to approach AI-generated content with informed caution—especially in academic, publishing, and public-facing contexts.

Ongoing Updates and AI Litigation

• U.S. Copyright Office:
  • Copyright and Artificial Intelligence Resource Page
  • Reports on AI and Copyright:
    • Part I: Digital Replicas
    • Part II: Copyrightability of Generative AI Outputs
    • Forthcoming Parts: Training data, licensing, and liability
• Congressional Research Service:
  Backgrounder on Generative AI and Copyright Law
• Financial Times, “Generative AI exists because of the transformer” by Visual Storytelling Team and Madhumita Murgia

Note: This is a rapidly changing legal area. Stay informed with updates from:

• U.S. Copyright Office
• Court rulings
• Legislative developments

Need Help? Contact a Copyright Librarian

Navigating copyright law—especially in the context of artificial intelligence—can be complex. If you have questions about:

• Whether your AI-assisted work qualifies for copyright protection

• How to responsibly use generative AI in academic or creative projects

• Citing or using copyrighted materials in AI training or outputs

• Legal and ethical considerations around AI-generated content

You don’t have to go it alone.

📚 Reach out to your campus copyright librarian or scholarly communications office.
They can help you understand your rights, responsibilities, and best practices for working with AI in a legally sound and ethically informed way.

Contact Information:

Sierra Whitfield, Copyright & Fair Use Librarian
Email: swhitfield@tamu.edu
Phone: (979) - 862 - 5438