A K-pop-based Explanation of Music Copyright Part 7.

Posted by Joseph Hwang Joseph Hwang 8/18/2024

A K-pop-based Explanation of Music Copyright Part 7.

Joseph Hwang

1. Generative AI (GPT) and Copyright

Artificial Intelligence has been around for a long time, invisibly affecting humans, but has recently become problematic. AI's ability to think, judge, and exercise its intentions can confuse humans. Still, in the copyright field, it has raised big questions about the fundamental nature of copyright.

The problem with AI is that it behaves like humans, or better than humans. This phenomenon has led to questions about human behavior, attitudes, and even the existence of self-consciousness, and essentially questions about morality. The confusion and problems AI has caused in the copyright field are not simply due to the creation of "human-like" AI. It all stems from "generative" AI.

By "generative" I mean data that is meaningful to humans also means that it is equally, perhaps more efficiently, capable of performing creative acts that only humans can do.

As explained previously, copyright is a uniquely human right. You don't get copyright just because you create, or even if you will. Copyright is recognized only when a human being produces it. This principle is the same in both Anglo-American and Civil Law. However, a third entity has appeared in this universe, which can do like humans.

The American lawsuit "The Naruto Case" became an iconic example of this principle of personhood. In 2011, an American photographer named David Slater was traveling through the jungles of Indonesia. During a moment of distraction, a monkey named Naruto snatched his camera and pressed the shutter several times, taking selfies of the monkey. David Slater published some of the photos in a book, and an animal rights organization called PETA (People for the Ethical Treatment of Animals) sued him for standing up for Naruto's friend. The court dismissed PETA's claim, saying that Naruto the monkey was not a person and therefore could not be sued. PETA appealed, and the U.S. Court of Appeals for the Ninth Circuit granted the appeal because it would be helpful for lower courts to address this developing area of the law. In its April 23, 2018 ruling, the Ninth Circuit held that PETA did not have standing as a friend of the monkeys and that animals do not have standing unless an act of Congress expressly grants them standing.

Although PETA lost the case, it is significant that it made the AI argument in its appellate brief. Animals cannot be recognized as litigants because they are creatures without personhood. This raises the question of how to legally protect the rights of impersonal generative AI and its products. The argument is that while the behavior of animals is not subject to human will, the existence of generative AI is the result of human creative activity, so if AI is not legally protected, then human creative activity, which is legally protected, is also vulnerable to infringement.

2. Deep learning and copyright

For generative AI to work, it needs to be preceded by mechanical learning. This is where deep learning comes in. The learning materials for deep learning are demanded such as human-generated text, video, and music files. This is mechanical crawling, which can lead to copyright infringement issues.

Musical compositions are created by placing and combining notes on a predetermined scale to create melodies and chords. Mathematically, the amount of music is limited to a fixed amount. If a generative AI learns how music is made by studying human music and creates new music based on that prior learning, it is no different from human composers. However, if it learns from the unauthorized crawling of data from websites, the crawling occurs already infringing copyright. Copyright law recognizes the exclusive rights of copyright holders to crawl the web. As AI learns more data, the royalties for using that data become prohibitively expensive. Developers say that improving AI requires a lot of investment and is delicate.

But if you think about it, humans learn music by listening to other people's work. Of course, when you buy a commercial record and listen to it, you're paying royalties to the copyright holders, but more often than not, we learn by listening to music for almost free.

3. Conclusion

No court has yet ruled on copyright infringement for musical works created by AI. Based on the principles and tendencies established in the cases decided by the courts, it is likely that musical works created by AI without personality will not be recognized as copyrightable. However, depending on the process of music creation by AI and the technology used, some copyrights may be recognized depending on the degree or extent of human intervention in the creation. Of course, it may be challenging to clearly distinguish the criteria and methods for determining the degree and scope of copyright recognition. Regardless, the copyright system will undoubtedly change as new perspectives and concepts on the creation of generative AI emerge. Indeed, copyright concepts have changed as technology and history have evolved.

* References and quotations:

https://www.wakeforestlawreview.com/2020/02/naruto-v-slater-one-small-step-for-a-monkey-one-giant-lawsuit-for-animal-kind/

Series Articles

Part 1.

https://www.musicbusiness.co.kr/2024/08/a-k-pop-based-explanation-of-music.html

Part 2.

https://www.musicbusiness.co.kr/2024/08/a-k-pop-based-explanation-of-music_01103635833.html

Part 3.

https://www.musicbusiness.co.kr/2024/08/a-k-pop-based-explanation-of-music_01024960771.html

Part 4.

https://www.musicbusiness.co.kr/2024/08/a-k-pop-based-explanation-of-music_01817572733.html

Part 5.