Generative AI & Its Use in Universities

There are many thorny ethical issues involved when generative artificial intelligence is used to create artistic works, including music. For an exploration of what is at stake, these three articles are a good starting point:

• Chow, Andrew R. “AI’s Influence on Music Is Raising Some Difficult Questions.” Time, December 4, 2023, https://time.com/6340294/ai-transform-music-2023/.
• Shroff, Lila. “AI & Copyright: A Case Study of the Music Industry.” GRACE: Global Review of AI Community Ethics vol. 2 no. 4 (2024), https://ojs.stanford.edu/ojs/index.php/grace/article/view/3226. 
• Smith-Muller, Talia. “AI Music: What Musicians Need to Know.” takenote: Inspiration for Music Makers. Berklee Online, August 21, 2023, https://online.berklee.edu/takenote/ai-music-what-musicians-need-to-know/.

An area of rapid expansion is generative tools to create and manipulate musical sounds. Each tool is subtly different, but the ones listed here (a partial list of tools in an ever-changing ecosystem) can be divided into two broad categories:

1. Tools that take written language (about genre, mood, speed, etc.) and translate it into musical sounds. These are designed to be accessible to any user, regardless of experience.
    
2. Tools that generate musical sounds, and then allow those sounds to be manipulated and shaped at various levels of detail / granularity using a visual interface. These are designed with more advanced users in mind, since they often require background knowledge in the use of digital audio workstations (DAWs).

Image Source: Culbreath, D. (n,d.). Generative AI and the future of music.  LinkedIn. https://www.linkedin.com/pulse/generative-ai-future-music-darren-culbreath-eaz5e/

DISCLAIMERS

1. Many of these tools make copyright claims about the contexts in which the music they generate can be legally used. Treat these claims with skepticism - this is a fundamentally unsettled area of copyright and intellectual property law.

2. Many of these tools make claims about the kinds of datasets that were used to train them, and the legal status of those datasets. Treat these claims with skepticism - it is very difficult to verify such claims.

TOOLS
Text-to-Music Sound Generators (generating sound based on written language cues).

Generative Digital Audio Workstations (manipulating and reconfiguring A.I. generated sounds)