Friday, March 24, 2023

Who owns the copyright in A.I.-generated works?

 

A.I.-generated works have been in the news for some time, and they are currently generating a lot of interest—both positive and negative. There are a lot of complicated issues with A.I.-generated works, and I am not going to discuss them all here at the moment, partially because the law is still evolving around them. However, there is one topic I briefly want to discuss: authorship and copyright protection.

In most instances, ownership of a copyrighted work initially belongs to the author of the work, and it is the owner of the copyright who can enforce those rights against others. If a work is not eligible for copyright protection, such as if it lacks creativity, is functional, or is in the public domain, then there is no copyright, and others cannot be prevented from using the work via copyright law.  

                In my opinion, one of the biggest issues that will determine how A.I.-generated works will impact creators is if copyright protection applies to these works. If an A.I.-generated work cannot be protected by copyright, then it will lack value for many companies. Without copyright protection, it would be incredibly difficult, if not impossible, to protect the work from use by others.

                When asked about this previously, my response has been that, based on the law as it stands, A.I.-generated works are not protected by copyright. Recently, the Copyright Office has issued two notices that provide further insight into how A.I.-generated works will be viewed.

                In the first memo, the Copyright Office clarified how copyright ownership would be applied to a graphic novel, Zarya of the Dawn, created using images from Midjourney.[1] The Copyright Office stated that Kristina Kashtanova would receive copyright protection in the text she wrote and the “selection, coordination, and arrangement of the Work’s written and visual elements.” The images generated by Midjourney, however, would not be protected by copyright because they were not created by a human. Essentially, Kashtanova owns the rights to the text and the overall appearance of the graphic novel, but not to the underlying images.

                In the second, the Copyright Office is proposing how it would treat applications for copyright protection of works created using A.I. The Copyright Office has stated that the determination of copyright protection will vary on a case-by-case basis. However, works generated purely by A.I. without addition or revision by a human author will not be entitled to copyright protection, and for works where a human author has used A.I.-generated work to create something else, the human author would receive copyright protection for the “human-authored aspects of the work…” or for the entire work if it is sufficiently creative enough to constitute “an original work of authorship.”[2]

                These two pieces appear to be in line with the current state of U.S. copyright law. A copyrighted work must be created by a human, but the amount of creativity needed to obtain a thin amount of copyright protection can be minimal. For instance, I can take a selection of public domain works, compile them into a book, and I would likely receive copyright protection for the book based upon the selection and arrangement of the works. Others are free to use the works independently, but using the same public domain works in the same arrangement as I selected could be a violation of my copyright.

                Fighting over authorship and copyright ownership is likely to be a major topic going forward. In order for A.I.-generated works to have true value, akin to traditionally human-authored works, then some form of copyright protection will be needed. The current state of the law affords A.I.-generated work minimal or no protection. The only way this will change is through litigation, and I find it hard to believe courts will diverge too far from the Copyright Office’s current interpretation of the law, or through lobbying to change the law.

                Personally, I would prefer a bright line rule of law prohibiting protection for A.I.-generated works. The primary reason for my view is because this technology will likely result in the loss of jobs for countless artists, writers, and other creatives. Second, I just don’t see the benefit of rewarding these types of works with copyright protection as they don’t really have artistic merit, and finally, because if A.I.-generated works are entitled to protection, the amount of litigation or threatened litigation that would ensue would be astronomical due to the ease with which A.I.-generated works are created and the amount of works that will be generated.

                Only time will tell how these issues will resolve, but I will be closely following the law as applied to A.I.-generated works very closely.  



[1] https://copyright.gov/docs/zarya-of-the-dawn.pdf

[2] https://www.federalregister.gov/documents/2023/03/16/2023-05321/copyright-registration-guidance-works-containing-material-generated-by-artificial-intelligence