AI or SI – What’s in a Name?

Before Led Zeppelin’s " Lemon Song appeared on Led Zeppelin II in 1969, the band had been performing an earlier version on stage. The band at that time called the song Killing Floor, named after the blues classic written and recorded by Chester Burnett, better known as Howlin’ Wolf.

As often happens when creatives work with existing material, the song evolved as Led Zeppelin continued to perform it. The band changed the arrangement, added new material, and gave it the unmistakable sound that made it sound like Led Zeppelin.

But Howlin’ Wolf’s music publisher filed a lawsuit against the band, claiming The Lemon Song infringed its copyright in Killing Floor. The dispute centered on how much of Howlin’ Wolf’s protected expression remained in the finished song. The lawsuit was settled, and later releases of the band’s album added Burnett as a songwriter.

However, the lawsuit highlighted an issue that copyright lawyers have had to deal with for decades. A songwriter can hear another song and take inspiration from it, and somewhere along the way, inspiration can become copying of protected expression. But the law doesn’t provide a precise percentage that determines when a work is completely new. The analysis is complicated and often requires expert analysis to decipher.

Generative artificial intelligence (AI) has complicated this age-old question. The US Copyright Office has taken the position that AI-generated works are not copyrightable. This position is partially based on the language in Article I, Section 8 of the US Constitution (those specifics are beyond the scope of this article.

The Copyright Office acknowledges that human contributions to work that’s partially AI-generated might be copyrightable. So for now, lawyers must evaluate whether there is sufficient human contribution for the work to be at least partially copyrightable. There’s no “bright-line” test or easy answer to that question; it requires that an attorney evaluate each work on a case-by-case basis.

On September 29, President Trump issued an Executive Order (EO) that may further complicate this legal question. The EO directs executive departments and agencies to replace “Artificial Intelligence” and “AI” with “Super Intelligence” and “SI” in official communications and other non-statutory materials. The EO doesn’t change statutes. Since the Copyright Office sits within the Library of Congress, not the Executive Branch, the EO also doesn’t impact the Copyright Office’s position. This article doesn’t take a political position on terminology, so I’ll use the term AI/SI.

The psychology behind the EO’s change in terminology made me wonder whether it might portend a change in the government’s approach to AI/SI. The EO describes AI/SI as technology that can “amplify human ingenuity” and “unlock new forms of creativity,” leaving open the possibility that AI/SI could be viewed as an extension of human creativity. And if AI/SI-generated works were viewed as human creativity rather than as something “non-human,” they could fall under the Copyright Office’s interpretation of US copyright law and Article I, Section 8. This article discusses the issue of copyright in AI/SI-generated works and explores the extent of human contribution required under current US law for copyrightability.

Who’s the Creator?

Until AI/SI, technological advancements weren’t considered to be obstacles to human creativity. For instance, if I put a pickup on my violin, the sound I produce can be amplified, distorted, looped, or processed with effects that would be impossible for me to achieve acoustically. But as the violinist, I would choose which notes to play, as well as the phrasing, articulation, dynamics, bowing technique, vibrato amplitude and speed, and countless other creative choices that distinguish one violinist’s performance from another’s.

A pickup can’t substitute for a violinist. But AI/SI can replicate the sound a violinist makes, and, coupled with a robot, AI/SI can probably play the violin, too. That distinction is why, unlike previous technological advances, AI/SI significantly complicates the analysis of whether a work is copyrightable.

Suppose a songwriter gives AI/SI the following prompt: “Write an upbeat pop song about unexpectedly meeting an old friend on a train, give it an eighties sound, and add a catchy chorus.” The system supplies the lyrics, melody, harmony, rhythm, instrumentation, and structure, and the user releases the first result. The human supplied an idea and some instructions, but under Copyright Office guidance, the finished work wouldn’t be copyrightable, since deciding the AI/SI prompt alone usually doesn’t provide sufficient human control over the expressive elements of the work such that the human is an author.

But suppose instead, the songwriter writes the song’s lyrics and asks AI/SI to generate a melody and harmony in the key of C to create a pop song with an eighties sound. AI/SI generates a melody, harmony, instrumentation, and changes the lyrics to better fit the new melodic line. Then, the songwriter asks AI/SI to change a bass line or the instrumentation and perhaps tweaks the melodic line and changes the harmony to work with the new melody. It becomes difficult to determine who the creator of the finished work is – and whether there is sufficient human input (by Copyright Office standards) for the work to have any copyright protection.

Suppose after viewing the AI/SI output, the songwriter manually makes numerous edits to the melody, harmony, and lyrics. The final product sounds more like folk music than eighties pop. This strategy is more likely to have involved sufficient human creativity for the song to be copyrightable, but only to the extent of the songwriter’s contributions. And like joint works with multiple collaborators, it can be challenging to determine which content was generated by the songwriter and which was generated by the AI/SI.

Businesses frequently counter similar issues. Suppose a company licenses a stock photograph of an outdoor seating area for use in an advertisement. The company has permission to use the photograph under the stock license, but it does not own the copyright in the original image. Before AI/SI, it was common for human creators to edit stock photos to meet the client’s needs. In those instances, the creator or their employer could own the copyright in the modified photo (subject to the license in the original).

Suppose instead, the human asks AI/SI to brighten the photograph, correct the perspective, remove a trash can and replace it with a more attractive café table, change the sky from gray to blue, and add landscaping. Based on the Copyright Office’s position, supplying prompts alone is unlikely to produce a copyrightable product.

Now suppose something closer to the Copyright Office’s example in its guidance. Suppose an individual has made a pencil sketch of an outdoor café. They’d like it to look like a photo for marketing purposes, so they ask AI/SI to modify the sketch so it appears to be a color photograph. Although the pencil sketch would be copyrightable, the AI/SI generated “photo” wouldn’t be.

What’s In a Name?

Words matter. Terminology that historically was acceptable now is disfavored because of the negative implications of words that were considered racist, sexist, or otherwise offensive. The term “artificial intelligence” hasn’t been tagged as offensive, but if the change to “super Intelligence” takes hold, it could change the way we think about AI/SI.

The EO expresses an intention to do just that: to transform how we think about AI/SI:

[AI] increasingly amplify human ingenuity and unlock new forms of creativity, empowering Americans to achieve what was previously impossible across science, medicine, and nearly every other domain of human endeavor. As these capabilities continue to improve, they increasingly represent not merely artificial intelligence, but a new era of Super Intelligence.. . . [T] he term “Super Intelligence” more appropriately captures the promise, potential, and rapidly advancing capabilities of these technologies.

Although the EO doesn’t change US Copyright Law, it does create a new paradigm under which to examine AI/SI-generated works and the extent to which they involve human creativity. And this new paradigm might change how creators approach AI/SI by encouraging them to use AI/SI to amplify their works, rather than as a substitute for human creativity.

Conclusion

The Lemon Song dispute centered on the extent to which Howlin’ Wolf’s protected creativity remained after Led Zeppelin modified the song. With AI/SI-generated works, the question is similar: How much human creativity was involved in creating the work?

It’s possible that the terminology in the recent EO might change how we think about AI and instead of allowing it to think for us, we use it to accelerate our own abilities. However, until that happens, the Copyright Office’s position remains: AI/SI-generated content isn’t copyrightable.

© 2026 by Elizabeth A. Whitman

Any references to clients and their legal situations have been modified to protect client confidentiality.

DISCLAIMER: The content of this blog is for informational purposes only and does not provide legal advice. No one should take any action regarding the information in this blog without first seeking the advice of an attorney. Neither reading this blog nor communication with Whitman Legal Solutions, LLC or Elizabeth A. Whitman creates an attorney-client relationship. No attorney-client relationship will exist with Whitman Legal Solutions, LLC or any attorney affiliated with it unless a written contract is signed by all parties.