Artist places black models in nature to create art to contemplate
The US Copyright Office considers an AI-generated work copyrightable if a human can prove that they themselves put a significant amount of creative effort into the final content, according to a policy published Thursday.
AI software capable of automatically generating images or text based on an input prompt or instruction has made it easier for humans to produce content. Accordingly, the USCO has received an increasing number of applications to register copyright protection for material, particularly artwork, created using such tools.
United States law states that intellectual property can only be copyrighted if it is the product of human creativity, and the USCO currently only recognizes work written by humans. Machines and generative AI algorithms therefore cannot be authors and their output is not copyrighted.
Digital art, poems, and books generated using tools such as DALL-E, Stable Diffusion, Midjourney, ChatGPT, or even the recently released GPT-4 are not copyrightable if they are human-created with only a text description or prompt. USCO director Shira Perlmutter warned.
“If the traditional elements of a work’s authorship are machine-produced, the work lacks human authorship and the Office will not register it,” she wrote in a copyright guidelines document.
“For example, if an AI technology only receives a prompt from a human and produces complex written, visual, or musical works in response, the ‘traditional elements of authorship’ are determined and executed by the technology – not the human user.
“Instead, these prompts work more like instructions to an artist on assignment — they specify what the prompter wants to display, but the machine decides how to implement those instructions in the output.”
The USCO is considering content created using AI if a human author has created something beyond the direct output of the machine. A digital artwork formed from a prompt and then further edited using Photoshop, for example, is more likely to be accepted by the office. The first image created using AI would not be copyrighted, but the final product produced by the artist may be.
So it seems that the USCO is simply saying, yes, if you use an AI-powered application to help create something, you have a reasonable chance of getting copyrighted, just as if you were using non-AI software. If it’s purely machine-made from a prompt, you’ll need to put a little more human effort into it.
In a recent case, officials registered a copyright certificate for a graphic novel containing images created with Midjourney. The overall composition and words were copyrighted because they were selected and arranged by a human, but the individual images themselves were not.
“In the case of works containing AI-generated material, the Bureau will consider whether the AI contributions are the result of ‘mechanical reproduction’ or rather than an author’s own original mental conception, to which [the author] visible form’. The answer depends on the circumstances, especially how the AI tool works and how it was used to create the final work. This is necessarily a case-by-case investigation,” the USCO said.
Perlmutter urged people applying for copyright protection for AI-generated material to clearly indicate how the software was used to create the content and to show which parts of the work were human-created. If they fail to accurately disclose this information, or attempt to hide the fact that it was AI-generated, USCO will cancel their registration certificate and their work may not be protected by copyright law. ®