Copyright, AI, and How the Laws Might Change

A recent appeals court decision, Thaler v. Perlmutter, reaffirmed that under current U.S. copyright law, only human-created works can receive copyright protection. The court upheld the Copyright Office’s longstanding position that human authorship is a fundamental requirement for copyright eligibility. While AI may be capable of generating creative works, the law does not currently recognize AI as an author.

How AI-Assisted Works Can Be Copyrighted

This ruling does not mean that AI-assisted works are automatically ineligible for copyright. Courts might find, consistent with existing law, that a human creates a copyrightable work when exercising creative judgment in selecting, arranging, or modifying AI-generated outputs. For example:

• An artist using an AI image generator like Midjourney might generate hundreds of images, refine prompts, adjust outputs, and manually edit the final composition.
• A writer working with ChatGPT might structure an entire narrative, develop characters, and selectively incorporate AI-generated text in a way that reflects their own creative input.
• A musician might use AI-generated melodies but rearrange them, add lyrics, and refine them into a final composition.
• A multimedia artist might use AI image creators along with large language models to create an illustrated, connected narrative. Even if no individual component is copyrightable, selection, coordination, and arrangement of works can be copyrightable, even if the underlying works are not. See, e.g., Feist Publications, Inc. v. Rural Telephone Service Co., Inc., 499 U.S. 340 (1991) (“[T]he statute dictates that the principal focus should be on whether the selection, coordination, and arrangement are sufficiently original to merit protection.”).

In these cases, the human’s role in coordinating and shaping AI output could potentially meet the threshold for copyright protection. The key is demonstrating that the AI was a tool under human direction allowing the exercise of sufficient “originality.”

The Problem of Ownership as AI Becomes More Advanced

As AI continues to improve, I predict that humans will use AI to generate increasingly complex content—including more complicated code, music, and entire works of literature—with minimal human intervention. This raises a critical question: Who owns AI-generated content if no human is considered the author?

You might be OK with nobody owning AI-based systems. However, even AI-created works, to be commercially valuable, will probably involve some level of investment. Consider the underlying purpose of the Copyright Clause, Article I, Section 8, Clause 8, of the Constitution, which says that, in order to promote the progress of science and the useful arts, works of authorship must be protected. In other words – we give people rights in what they create to give them an incentive to create the works and invest in their distribution. Doesn’t this same theory apply to AI-generated works? Do we suddenly NOT want works to be created and distributed simply because they are AI generated? Isn’t the world a better place if more software, more music, more art is available?

One particularly thorny issue arises in software development. If AI-generated code is not protected by copyright because no human “authored” it, companies that invest in AI-driven software development may struggle to secure legal protections for their innovations. For example, if code is created entirely by AI and a company licenses that software to its customer, then what is to stop the anyone from getting a copy and making their own business competing with the original? Businesses might choose to avoid entering the market in the first place, and the world would be deprived… of what could have been!

A Future Shift Toward AI Copyright Protection?

Looking ahead, economic and business interests may drive changes in copyright law. As AI becomes more integral to creative industries, there could be increasing pressure to extend copyright protection to certain AI-generated works. Possible legal changes could include:

• Granting copyright to the human who directed the AI in a sufficiently creative way.
• Recognizing AI as an “assistive” author, similar to how corporate works are sometimes treated under work-for-hire rules.
• Creating a new category of intellectual property rights for AI-generated content, distinct from traditional copyright.

Over time, copyright law may evolve to reflect the reality that AI is not merely a tool but an active participant in the creative process. Whether this shift happens through court decisions, legislative updates, or industry-driven solutions remains to be seen—but one thing is certain: the role of AI in creative work is only going to grow, and the law will need to adapt accordingly.

But this isn’t just my theory about what might happen. It’s applying what’s happened in the past to predict what might happen in the future. For example, the Copyright Act of 1909 didn’t cover things like sound recordings and computer software (in part because they didn’t exist at the time). Commercial realities changed, and the law changed to reflect that.

The Sound Recording Act of 1971 expressly modified the Copyright Act of 1909 to protect sound recordings made on or after February 15, 1972. Then, Congress got around to passing the Copyright Act of 1976, which attempted to make clear that original works should be protected, regardless of what new media was developed, by stating that, “Copyright protection subsists, in accordance with this title, in original works of authorship fixed in any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.” See 17 USC 102.

But that still wasn’t enough, because there was still an open question about whether the law covered computer programs due to the nature that, well, they’re highly functional (and copyright law generally protects artistic things, not functional things). Starting around 1974, the National Commission on New Technological Used of Copyrighted Works (CONTU) researched computer software issues and issued a final report in 1978 suggesting the ways in which the law might adapt to technology. Then, Congress passed the Computer Software Act of 1980, allowing the protection of computer software as a literary work (software is still registered using Form TX), protecting the written code but not the functionality or ideas that the code implemented.

It wouldn’t surprise me if history repeats itself, or at least rhymes.