Tools Do Not Create: Human Authorship in the Use of Generative Artificial Intelligence

30 Pages Posted: 13 Jul 2023 Last revised: 29 Feb 2024

See all articles by Michael D. Murray

Michael D. Murray

University of Kentucky, J. David Rosenberg College of Law

Date Written: February 27, 2024


· Tools Do Not Create: Human Authorship in the Use of Generative Artificial Intelligence, 15 CASE W. RESERVE J.L. TECH. & INTERNET 76 (2024),
Artistic tools do not create paintings and drawings. Brushes, paint, pencils and pastels, video and photographic cameras, image editing tools such as Adobe Photoshop, and ever increasingly complicated algorithms in neural networks, foundation models, and large language models are not the authors of artworks. Human artists create content. Human artists use tools to create visual art.

This sketch of the process of creation of art has been blurred in recent months by the advent and rapid adoption of visual generative artificial intelligence (AI) tools such as DALL-E 2, Stable Diffusion, and Midjourney that inspire magical thinking regarding the creation of artworks. The very name “generative” AI suggests a narrative that the algorithms, programming, foundation models, and transformer technology composing these AI tools are the actual authors of the works produced—that the artworks are “created” in the copyright sense of the word by the AI. Not so.

The United States Copyright Office has recently issued guidance on the copyrightability of visual works that artists and authors have produced using visual generative AI tools that might be limited in their copyrightability because the works or elements of the works were generated by the AI. Thus, the Copyright office has taken the bait and swallowed the narrative that the magic box of generative AI tools actually performs the steps of authorship of original and creative copyrightable works: the AI somehow “conceives” of the image in its “mind” and somehow “randomly” or “automatically” renders it into existence in a fixed and tangible medium. The radical core of the Copyright Office’s interpretation is sound: randomly or automatically generated works do not have human authorship; they are not conceived of in the minds of human authors and the human authors do not cause their inner conceptions and designs to be rendered into fixed and tangible forms. The error comes in the Copyright Office’s thinking that generative AI systems randomly or automatically create and generate works.

Contemporary visual generative AI systems can do extraordinary things, but as of yet not autonomously and not automatically. It is a fallacy to view AI systems as the authors of the works they generate. The process of how an end-user of a contemporary generative AI tool creates art and how a human artist goes about the same task are very similar.

Generative AI systems are tools—highly complex, deeply technological tools to be sure, but tools none the less. And these tools require a human author or artist—the end-user of the generative AI system—to provide the inspiration and design and often the instructions and directions on how to produce the image. An artist working with a generative AI tool is no different from an artist working with a digital or analog camera or with Photoshop or another image editing and image rendering tool.

Keywords: Generative AI, artificial intelligence, authorship, copyright, originality, creativity, human authorship, DALL-E 2, Stable Diffusion, Midjourney, Photoshop, artistic process

JEL Classification: K1, K10, K3, K30, K40, K41

Suggested Citation

Murray, Michael D., Tools Do Not Create: Human Authorship in the Use of Generative Artificial Intelligence (February 27, 2024). 15 Case Western Reserve Journal of Law, Technology & the Internet 76 (2024), Available at SSRN: or

Michael D. Murray (Contact Author)

University of Kentucky, J. David Rosenberg College of Law ( email )

620 S. Limestone Street
Lexington, KY 40506-0048
United States
219-299-9777 (Phone)
859-323-1061 (Fax)


Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics