Creativity Without a Creator? The Copyright Problem of AI

By Estefania Cerna '28

---

Who Owns Creativity in the Age of AI? 

Imagine typing a single sentence into an artificial intelligence (AI) tool and receiving a fully formed painting, song, or essay in seconds. No artist. No writer. No human hand shaping the final product. Just code. Now ask the question copyright law was never designed to answer: Who owns it? 

U.S. copyright law has historically required human authorship to qualify for protection. (1) Generative AI challenges that rule by producing creative works without being legally recognized as an author. (2) The result is a growing legal gray area that existing law is not fully equipped to resolve. To address this gap, copyright law must evolve to recognize AI-assisted works and clarify the level of human involvement required for authorship.

Copyright Was Built for Humans 

Copyright law in the United States originated as a system designed to protect human creators. (3) The Copyright Act of 1790 was the first federal copyright statute and granted protection to authors of books, maps, and charts. (4) Its purpose was to incentivize individual creativity by granting exclusive rights to creators. (5) The Copyright Act of 1909 expanded those protections by addressing reproduction and distribution rights and clarifying the public domain. (6) Lastly, the Copyright Act of 1976 modernized copyright law and extended protection to “original works of authorship” across a wide range of mediums. (7) Throughout these developments, copyright law has consistently centered on human creativity. (8) 

Copyright protection requires both originality and authorship. (9) Originality requires a minimal degree of creativity rather than novelty or effort alone. (10) In Feist Publications, Inc. v. Rural Telephone Service Co., the Supreme Court clarified that copyright law protects original expression, not simply the effort or labor involved in creating something. The case involved a telephone directory, where the Court found that listing names and numbers alphabetically lacked the minimal creativity required for copyright protection. This decision established that originality requires at least a small degree of creative choice, not just hard work. (11)

Courts have also consistently limited authorship to humans. In Naruto v. Slater, a case often referred to as the “monkey selfie” case, a macaque named Naruto took photographs using a photographer’s camera. An animal rights group argued that the monkey should own the copyright to the images. The court rejected this argument, holding that non-human entities cannot be recognized as authors under U.S. copyright law. (13)

Generative AI systems can produce writing, art, and music that appear original. (15) Platforms like Midjourney are generative artificial intelligence programs that create images from written prompts using diffusion models, a type of machine learning that transforms noise into detailed visuals based on patterns learned from vast datasets of existing images. Users interact with Midjourney by entering descriptive text commands, and the system generates multiple visual interpretations, often resembling styles found in human-created art. (16) However, AI systems cannot qualify as authors because they are not human. (17) If a work is generated entirely by AI without meaningful human input, it may not qualify for copyright protection. (18)

The law remains unclear on how much human input is required when a person prompts an AI system. (19) This places copyright law in a difficult position, where AI can generate a work but cannot claim ownership, while the human contribution may be too minimal to qualify for authorship.

Is AI Training Legal or Infringement? 

Generative AI systems are trained on large datasets that may include copyrighted works. (20) This raises the legal question of whether such use qualifies as fair use under copyright law. (21) In Authors Guild v. Google, Inc., Google scanned millions of copyrighted books to create a searchable database known as Google Books. Although full copies of the books were made, the court held that Google’s use was “transformative” because it did not provide the books themselves to the public, but instead allowed users to search for terms and view limited snippets of text. The court emphasized that this new function served a different purpose than the original works and provided significant public benefit, ultimately qualifying as fair use. (22) AI developers argue that training models are similarly transformative because copyrighted material is used to analyze patterns, styles, and relationships rather than to reproduce or distribute the original works. In this way, the content is converted into functional data that helps generate new outputs, supporting the claim that the use serves a distinct and transformative purpose. (23)

However, AI systems generate outputs that may resemble original works, which complicates the fair use analysis. (24) Fair use is evaluated under a four-factor test that considers the purpose and character of the use, including whether it is transformative; the nature of the copyrighted work; the amount and substantiality of the portion used; and the effect of the use on the potential market for the original work. Courts must weigh these factors together, rather than relying on any single one. When applied to AI, this analysis becomes more complex because training may be considered transformative, but the outputs can sometimes closely resemble original works and potentially compete with them in the market. Courts have not yet definitively ruled on whether AI training qualifies as fair use. (25)

Major entertainment companies, including The Walt Disney Company and Universal Pictures, have initiated legal action involving AI platforms such as Midjourney. (26) These cases reflect broader efforts to apply existing copyright law to emerging AI technologies. (27) The outcomes of these disputes may significantly influence how copyright law addresses AI-generated works and training practices. (28) If AI-generated works are not protected, they may enter the public domain immediately. (29) At the same time, copyrighted works may be used to train AI systems without compensation to their creators. (30) This raises concerns about fairness, economic impact, and the future of creative industries. 

Where Do We Go From Here?

Congress needs to clarify what level of human involvement qualifies as authorship. However, copyright law should evolve to reflect how creativity actually works today. AI is not replacing human creativity, but expanding it, and tools like generative AI function more like advanced instruments than independent creators. Just as cameras and digital editing tools changed creative processes in the past, AI still depends on human direction, input, and decision-making. Because of this, the law should recognize AI-assisted works as eligible for copyright, even when the human role is less traditional than it has been in the past.

A key challenge is determining where to draw the line between AI-assisted and fully AI-generated works. Not every use of AI should qualify for protection. When a user simply enters a short prompt and accepts the output without modification, the human contribution may be too minimal to meet the standard of authorship. In contrast, when a user engages in a more involved creative process, such as refining prompts, selecting among outputs, editing, or combining AI-generated material with original human expression, the result reflects meaningful human authorship. The law should focus on whether the human exercised creative control over the final work, rather than on the mere use of AI as a tool.

For example, Congress could adopt a standard granting copyright protection to AI-assisted works where a human makes substantial creative decisions that shape the final output. This could include selecting or arranging AI-generated elements, editing or modifying the content, or directing the AI through iterative prompting that reflects original creative judgment. At the same time, works generated entirely by AI with minimal human input would not qualify for protection and would remain in the public domain. This type of framework would provide clearer guidance while preserving the core principle that copyright protects human creativity.

Limiting protection too strictly risks slowing innovation and ignoring the reality that technology has always shaped creative expression. Instead of relying only on restrictive approaches like licensing frameworks or heightened regulation, copyright law should adapt by allowing broader recognition of human involvement and treating AI training as largely transformative. By doing so, the law can better support both creativity and technological progress, rather than forcing them into conflict.

Endnotes

1. U.S. Copyright Office, Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence (2023). https://www.federalregister.gov/documents/2023/03/16/2023-05321/copyright-registration-guidance-works-containing-material-generated-by-artificial-intelligence

2. Ibid.

3. Copyright Act of 1790, ch. 15, 1 Stat. 124. https://www.copyright.gov/history/1790act.pdf

4. Ibid.

5. Ibid.

6. Copyright Act of 1909, ch. 320, 35 Stat. 1075. https://www.copyright.gov/history/1909act.pdf

7. Copyright Act of 1976, 17 U.S.C. § 101 et seq. https://www.copyright.gov/title17/

8. Ibid.

9. 17 U.S.C. § 102(a). https://www.law.cornell.edu/uscode/text/17/102

10. Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991). https://supreme.justia.com/cases/federal/us/499/340/

11. Ibid.

12. U.S. Copyright Office, Compendium of U.S. Copyright Office Practices § 306. https://www.copyright.gov/comp3/

13. Naruto v. Slater, 888 F.3d 418 (9th Cir. 2018). https://cdn.ca9.uscourts.gov/datastore/opinions/2018/04/23/16-15469.pdf

14. Ibid.

15. U.S. Copyright Office, Artificial Intelligence and Copyright (2023). https://www.copyright.gov/ai/

16. Ibid.

17. Ibid.

18. Ibid.

19. Ibid.

20. Ibid.

21. 17 U.S.C. § 107. https://www.law.cornell.edu/uscode/text/17/107

22. Authors Guild v. Google, Inc., 804 F.3d 202 (2d Cir. 2015). https://law.justia.com/cases/federal/appellate-courts/ca2/13-4829/13-4829-2015-10-16.html

23. Ibid.

24. U.S. Copyright Office, Artificial Intelligence and Copyright (2023). https://www.copyright.gov/ai/

25. Ibid.

26. Bobby Allyn, Disney and Universal sue AI company Midjourney over copyright infringement, NPR (June 12, 2025), https://www.npr.org/2025/06/12/nx-s1-5431684/ai-disney-universal-midjourney-copyright-infringement-lawsuit.

27. Ibid.

28. Ibid.

29. U.S. Copyright Office, Artificial Intelligence Guidance (2023). https://www.copyright.gov/ai/

30. Ibid.