In the last couple of years, the world has seen the steep development curve of artificial intelligence, especially in the generative form. The transition from research to the commercial market takes place as we speak. A real-life example of this is the festival R², organised by Reinier Zonneveld (Veenhoven, 2024). He has developed AI software and played a back-to-back with his digital counterpart on the festival. The implementation of generative AI in creative industries such as music composition resulted in the creation of multiple tools. Examples of these are Suno AI and Udio AI. These tools allow the user to generate music with one simple text prompt.
A Dutch newspaper wrote an article about these apps, criticising that with these apps users are able to create songs in a matter of minutes, resulting in a flood of music that is well composed but misses the uniqueness a real artist can bring to the table. Famous artists feel threatened by these tools not only because of the competition they bring but also because of copyright infringement issues(Hijink, 2024).
While these applications use training data to develop their algorithms, it is often not allowed to use artist names in the prompt; however, with the right prompt query and a bit of luck, you can create similar songs to mimic the style of Taylor Swift by using descriptive words such as “pop, alternative folk, emotional, female vocals”(Hijink, 2024). In the United States, a bill is introduced in the US Congress that would require companies to share the training dataset with the Register of Copyrights before launcing tools with generative purposes. While its intention is not to ban training on copyrighted music completely, it should put a responsibility on these companies to provide information on the massive amounts of training data that is used and usually kept private(Robins-Early, 2024).
Personally, I think that these tools are awesome and that artists should not feel threatened by them but should embrace them in their music creation process. These tools can spark creativity and shorten the amount of time spent on components that can be replaced by AI, such as mastering the music. Another benefit of co-creating music with AI tools is that they can create different complementing music progressions that would otherwise be lost. Besides the process of music creation, I believe that people will prefer human-made music since music is a way of expressing emotions, and the experience of going to a concert of your favourite artist would not feel the same. Finally, for beginning artists and enthusiasts, these tools lower the barrier set to creating professional music.
References:
Hijink, M. (2024, May 17). Voor 2 cent per ‘liedje’ perst AI de popmuziek uit. NRC. https://www.nrc.nl/nieuws/2024/05/17/voor-2-cent-per-liedje-perst-ai-de-popmuziek-uit-a4199269
Robins-Early, N. (2024, April 9). New bill would force AI companies to reveal use of copyrighted art. The Guardian. https://www.theguardian.com/technology/2024/apr/09/artificial-intelligence-bill-copyright-art
Reinier Zonneveld, techno-dj. (2024, April 10). NPO Radio 1. https://www.nporadio1.nl/fragmenten/kunststof/9bc5c481-8048-4408-9e9b-99fd8befcbed/2024-04-10-reinier-zonneveld-techno-dj
