What is (copyright) protected Art?

by on 16. October 2023

The more diverse AI tools are being used for content creation, especially in the field of image generation, the more complex the discussion about the (copyright) protection of the AI-generated output becomes.

Reason enough, to take a closer look at the arguments for protection that have been put forward by authorities and courts to date. The following considerations are made in the context of US law and the legal systems of the European Union, both of which do not provide for the specific case of computer-generated protected works (unlike the regulation in the UK in Section 9 (3) of the Copyright, Designs, and Patents Act 1988).

The starting point is therefore that only the result of a personal, i.e. human, intellectual creation can enjoy copyright protection as a “work” – e.g., § 2 section 2 of the German Copyright Act (UrhG) and the U.S. counterpart 17 U.S.C. § 102 (a) in conjunction with the common interpretations of the courts, which have so far unanimously limited “works of authorship” to creations by human authors.

In particular, the U.S. Copyright Office has been considering the protectability of various AI-created content in recent months:

1. “Zarya of the Dawn” by Kristina Kashtanova
For example, in connection with the graphic novel “Zarya of the Dawn” by Kristina Kashtanova.
In this novel, Midjourney’s AI was used to create the images. The author created text prompts as “central creative input”, selected one or more images for further use, and then refined/modified the prompts to create the final image. The text, on the other hand, was created entirely without the aid of a generative AI program.

The U.S. Copyright Office stated in its decision from Feb. 21, 2023 that ” Ms. Kashtanova is the author of the Work’s text as well as the selection, coordination, and arrangement of the Work’s written and visual elements. That authorship is protected by copyright. […] the images in the Work that were generated by the Midjourney technology are not the product of human authorship.”

The Copyright Office denied copyright protection based on the following arguments:

  • Rather than a tool that Ms. Kashtanova controlled and guided to reach her desired image, Midjourney generates images in an unpredictable way.
  • The information in the prompt may “influence” generated image, but prompt text does not dictate a specific result.
  • The fact that Midjourney’s specific output cannot be predicted by users makes Midjourney different for copyright purposes than other tools used by artists.
  • Prompts function closer to suggestions than orders, similar to the situation of a client who hires an artist to create an image with general directions as to its contents, client would not be author of that image.

Conclusion: In particular, the criterion of the predictability of the result when using an AI tool is important for copyright protection.

2. “Théâtre D’Opéra Spatial” by Jason M. Allen
In its September 5, 2023 decision, the Copyright Office dealt with a much more human-driven use of generative AI: Jason M. Allen created an image using Midjourney but then made at least 624 revisions and optimizations to the text prompts afterwards (setting the scenery, selecting sections and determining the shading) before the image was finalized.

Despite these extensive revisions, the Copyright Office “[…] concludes that the Work contains an amount of AI-generated material that is more than de minimis and thus must be disclaimed. […] the Midjourney Image, which remains in substantial form in the final Work, is not the product of human authorship.

This was based on the following considerations:

  • Mr. Allen‘s sole contribution to the Midjourney Image was inputting the text prompt that produced it.
  • Although inputting at least 624 revisions and text promts, the steps in the process were ultimately dependent on how the Midjourney system processed Mr Allen‘s prompts.
  • Process of prompting can involce creativity, some prompts may be sufficiently creative to be protected by copyright, but that does not mean that providing text prompts to Midjourney actually forms the generated images, Mr. Allen‘s actions did not amount to creative control of the claimed elements of the work.

Conclusion: The human author must retain creative control over the creative process for the result to enjoy copyright protection.

3. “A Recent Entrance to Paradise” by Steven Thaler
Somewhat older in date is the work “A Recent Entrance to Paradise” by Steven Thaler, which was the subject of a judicial decision by the District Court of Colombia on August 18, 2023. Thaler created the image with his own AI algorithm, “Creativity Machine”, and the U.S. Copyright Office had already denied registration on February 14, 2022, arguing, “Because copyright law as codified in the 1976 Act requires human authorship, the Work cannot be registered”.

The court upheld this classification, stating, “The Copyright Office acted properly in denying copyright registration for a work created absent any human involvement.

This decision was based on the following criteria:

  • Copyright has never stretched so far, however, as to protect works generated by new forms of technology operating absent any guiding human hand.
  • Human authorship is a bedrock requirement of copyright.
  • But in this case the only question is whether a work generated autonomously by a computer system is eligible for copyright and in the absence of any human involvement in the creation of the work, the clear and straightforward answer is no.

At the same time, the court claimed: ” Undoubtedly, we are approaching new frontiers in copyright as artists put AI in their toolbox to be used in the generation of new visual and other artistic works.”

4. Transferability to German Law?
So far, there are no court decisions in Germany regarding the protection of “works” generated by AI. However, given that human creativity is the source of copyright protection, it seems that the criteria, used by U.S. authorities and courts, could be transferred and used for classification in Germany as well, leading to the same results.

5. Under what conditions could AI output be copyrighted?
According to the current state of the law, AI output is only copyrightable if and to the extent that the output results from a direct, human-controlled contribution. This is the case, for example, when a photo/image, created by a human, is “read” into the AI, to further develop the AI´s input. In this case, the “foundation” – the human-created image – is and remains protected as long as it remains recognizable in the AI´s output. The same applies, for example, if the AI output is further processed and/or developed by humans to a sufficient extent, e.g. if texts are created by means of a text generator, but then supplemented and rewritten to a sufficient extent. Attention: Only the human part is then protected by copyright, the unedited AI output remains excluded from copyright protection.

6. Future Prospects
Further development will certainly become more differentiated, as AI tools will be able to work “closer” to human specifications. This could mean that the criteria of predictability and control are met in individual cases and the AI tool could actually become a purely human-controlled tool. This is why, the discussion about the protection of AI output remains exciting – and partly reminiscent of the discussion of the rise of photography in the 19th century, when photographers were initially denied protection of their work comparable to that of visual artists, since the photographer only operated the camera’s shutter.