31 Jan Navigating A.I.: Understanding Ownership in Automated Content
The world of copyrights has undergone several major decisions during the past year by the U.S. Copyright Office (“USCO”), likely leaving people frustrated and helpless. In short, the USCO has found that artificial intelligence-generated images are unlikely to obtain federally granted copyright protections. Copyrights are considered to be original works fixed in a tangible medium of expression. See 17 U.S.C. § 102(a). Further, the work must be independently created by an author and possess sufficient creativity. See Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345 (1991).
The Supreme Court of the United States has found that the work must possess more than a de minimis quantum of creativity. With the baseline being that no copyright is issued in which “the creative spark is utterly lacking or so trivial as to be virtually nonexistent.” Id. at 359 and 363.
The following are different AI-generated works the USCO has made registration decisions. Providing insight for authors seeking to obtain federal copyright protections for AI-generated works:
Zarya of the Dawn by Kristina Kashtanova
Here, Ms. Kashtanova created a comic book and utilized a generative AI platform, known as MidJourney, to create a series of images which she then made minor changes to in Photoshop and then filed the work with the USCO. At first, the USCO granted federal protection, but upon discovering the author utilized generative AI to create the images, reversed its decision to the extent that the author was granted protection for the text of the comic book and the selection, coordination, and arrangement of the comic’s written and visual elements.
In short, the decision by USCO in Zarya of the Dawn has revealed several important points: (1) authors are required to disclose the use of AI generative platforms; (2) USCO does not grant federal copyright protections to works created by AI generative platforms unless the image is found to be sufficiently altered by the author to qualify for copyright protection; and (3) authors may use AI generative platforms to create works, and then the collection and assembling of such work as a compilation may be granted federal copyright protection if the material and data are selected, coordinated, or arranged in a sufficiently creative way.
Théâtre D’opéra Spatial by Jason M. Allen
Mr. Allen’s two-dimensional artwork was created through the use of MidJourney, an AI generative platform. Mr. Allen was unwilling to disclaim the use of the AI generative platform, claiming that he input numerous revisions and text prompts, at least 624 times, before producing an initial version of the image. He then further altered using Adobe Photoshop to remove flaws and create new visual content, and Gigapixel AI to enhance the image resolution and size.
Despite all of this, the USCO determined that the image created was not of human origin. USCO provided guidance that when considering an application for registration, USCO will ask whether the work is of human authorship. Whether by using a computer or device as an assisting tool. or whether the work was actually conceived and executed by a machine and not a man. See Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190, 16,192 (Mar. 16, 2023).
Ultimately, USCO found that the work by Mr. Allen remained similar to the AI-generated image. That the alterations made by him were not sufficient to make the work created by a human author. Further, USCO revealed that:
- Sufficient information regarding adjustments made by Adobe Photoshop or similar photo editing software must be provided to USCO.
- The use of Gigapixel AI or similar size and resolution-altering software does not alter or introduce new elements to an image and thus would not be enough on its own to transform an image to receive federal copyright protection.
- Where the use of an AI generative system is more than de minimis in the creation of a work, the content must be disclaimed in the application for registration.
Note: Mr. Allen intends to file a federal lawsuit regarding his application’s denial.
Suryast by Ankit Sahni
Mr. Sahni inputted an original photograph he authored and a copy of Vincent van Gogh’s The Starry Night into RAGHAV, an AI Painting App, where The Starry Night was to be used as the style applied to the original photograph. Mr. Sahni exerted no control over the placement, appearance, or colors, aside from the intensity of the elements. Rather, RAGHAV created the final rendering and was more than an assistive tool.
In the eyes of the USCO, Suryast is not protected because “protection is given only to the expression of the idea—not the idea itself.” Mazer v. Stein, 347 U.S. 201, 217 (1954). Here, Mr. Sahni contributes to an idea, while RAGHAV actually expresses the idea. Further, since RAGHAV is an AI Painting APP and not a human, a copyright may not be registered under it either. The only possible copyrightable work is the original photograph taken by Mr. Sahni.
Ultimately, USCO retains that (1) only humans may be the creator and register copyrights; (2) the creator must do more than just develop the idea, the creator must actually express the idea; and (3) work created by AI is not afforded copyright protection.
Ownership in Automated Content
Artists, authors, business owners, creators, and the like, beware and be cautioned when using AI applications. Specifically in the assistance of creating your work. Copyright protections are unlikely to be afforded if the AI is actually expressing the idea despite input put in by the creator. However, generative AI could be used to create trademarks (logos) for goods and services and potentially afforded protection. Subject to similar existing trademarks and acceptance by the United States Patent and Trademark Office. If you are interested in registering your copyright or trademark, or have further intellectual property needs, please contact us today.
Written by Vikesh N. Patel, Attorney at KM&D Law