The U.S. Copyright Office has released a series of highly anticipated reports that address the copyrightability of AI-generated content. The Office maintains that copyright protection is only available for works with human authorship—as rooted in the U.S. Constitution—but now acknowledges scenarios where AI-assisted works may qualify. Still, fundamental copyright principles apply.   

Whether a work has sufficient human authorship to qualify for copyright protection must be evaluated on a case-by-case basis.  And works created entirely by AI, like other machine and non-human sources, aren’t copyrightable because they lack basic human creative influence.  Whether AI tools are used to assist a creator versus a stand-in for human creativity is the key question in determining copyrightability. In a nutshell: the nature of the human author’s creative contribution is what matters.   

A human author is entitled to copyright protection where his creative contributions are perceptible in the output or when the author creatively enhances, selects, arranges, or modifies AI-generated content. The Office likened the analysis to copyright ownership for joint authors, where the focus is on the amount of control necessary to claim authorship. For example, after country singer Randy Travis suffered a stroke that affected his speech, he used an AI tool to enhance and modify his damaged voice to create new tracks. Because Mr. Travis contributed creatively by providing his voice in original songs and only used AI tools to assist in the modification and enhancement, his new songs were also copyrightable. The same is true for an artist who inputs her original painting and uses an AI tool to arrange new elements or modify the work. 

Based on the current state of generative AI, content created solely through the use of prompts is unlikely to be eligible for copyright protection. Current AI systems—sometimes referred to as “black boxes”—do not allow users to exert sufficient creative influence over the output. Users can repeatedly enter the same prompt again and again and the output may vary every time. This gap shows that a human author lacks control over how his/her idea is converted into a fixed expression. Additionally, under current systems, AI may fill in any aspects not prescribed by the author, changing, expanding, or even overriding the author’s creative expression and ideas. The Office noted that someday, generative AI may advance to allow users increased control over AI-generated outputs to reflect human creative contribution, which may change the copyrightability landscape again. The report also notes that AI prompts themselves may qualify for copyright protection if they are sufficiently creative.   

Reasoning its analysis, the Office cited the Copyright Act’s long history of flexibility with emerging technologies, such as cameras and computers. For instance, copyright protection was extended to photographs, video games, software, and motion pictures, all of which required an analysis of a human author’s role and contributions in the creative process. AI tools are no different.  For now, the Office has concluded that no new copyright laws or different protection for AI-generated works is necessary. Instead, current copyright law is able to adapt, as it has in the past, to accommodate today’s technological advancements.  

Key Takeaways: 

• AI-generated content is copyrightable so long as a human author’s creative input is reflected in the output or when the human author creatively modifies, arranges, selects, enhances, etc. AI-generated content. 

• Content generated from AI prompts alone is unlikely copyrightable because current generative AI systems do not allow sufficient human control over how the output is expressed. 

• New legislation, laws, or different protection for AI-generated works are unnecessary at this time.  

• Copyright applicants still need to identify and disclaim non-human expressions. 

To get in touch with Dianne directly, contact her here.