Who Gets the Copyright?

The following images are my new work for the fictitious fashion brand “Amora.” It’s a research study that I am working on intermittently. I began teaching myself AI promptgraphy to explore what can be accomplished in various timeframes and to understand the limitations. It's an addictive process with nearly endless opportunities, and with each iteration of the “program,” it becomes more realistic and refined.

In any creative industry, the immediate question is always, "What now? Will we exist in the near future, or do I need to acquire new skills to earn a living?" Fortunately, before AI-generated creative content becomes mainstream, two main issues need to be resolved. The first is the licensing of large content databases that AI uses to deep learn its “skills.”

To delve a bit deeper, let's consider my Midjourney prompt, which created these images: “A Patrick Demarchelier-style photo of a beautiful model posing with Christian Dior haute couture in front of the library of the ancient city of Ephesus at golden hour, ultrarealistic, shot with a Sony A7 9, 50mm f2.8 DSLR, --ar 103:45 --style raw --stylize 0 --weird 250.” As you can see from the prompt, AI has already learned about fashion photographer Patrick Demarchelier, understands Western beauty standards, recognizes the library of Ephesus and its ancient ruins, and also comprehends the concept of posing in golden hour and what Christian Dior haute couture looks like.

One of the largest content libraries with this kind of variety of images is Getty Images. Their content is licensed, and they conveniently create “keywords” for every detail in an image they represent. Naturally, it is one of the most desirable libraries for AI training. There is an ongoing lawsuit, "Getty Images vs. OpenAI," which will likely be one of the determining court cases for licensing guidelines.

The second issue, which I believe is more important for creatives, is who owns the copyright to AI-based content? The United States Copyright Office, a part of the Library of Congress, registers copyright claims and records information about copyright ownership. They have rejected multiple cases (these are handled on a case-by-case basis) of AI-generated content for copyright claims. The reason for their rejection is the 1976 Copyright Act, which requires a work to be fixed “by or under the authority of the author.” Here, I am referring to an article written by Robyn Bowland and Ariel Katz at Irwin IP law firm, who specialize in copyright law. They explain, “While the word ‘author’ is not defined in the Copyright Act, the text requires a copyrightable work to have an originator with the capacity for intellectual, creative, or artistic labor. The Court stated this means that the originator must be a human. Those seeking copyright protection over works that utilize AI should ensure that the human artist, not AI, controls the creative selection and arrangement elements of the work and that the human contribution of creative expression is more than de minimis while separately identifiable.”

In short, until Congress passes a law that specifically states that a human-prompted, AI bot-generated content is copyrightable, it is fair to say that any AI-originated content is fair use and not copyrightable. So please go ahead and copy my prompt or use any of the images for your projects.