Ramona Hernandez de Sant'Ana/The Mossy Log
Like many music lovers, I often attempt the Sisyphean task of making a full list of my favorite albums. While I do this, I look back on what brought me to these records and how they have become so essential to my life. In reflecting upon my music discovery habits over the years, I have come to the conclusion that it is necessary for a listener to be wary of — if not completely avoid — algorithms when engaging with music.
Specificity is necessary here. The term “algorithm,” like “artificial intelligence,” is often overused to the point of losing meaning. When I use the word algorithm, I am referring to music recommendation algorithms: automated systems used by a streaming service to show music to a consumer based on past listening habits. This category of algorithm encompasses everything from Spotify’s “Discover Weekly” playlist, album recommendations on a streaming service’s front page, and autoplay features that play similar tracks after a queue or album has concluded.
These algorithms present the listener with an infinite amount of songs they have never heard before, which sounds appealing on the surface. While they do deliver in this regard, they promote a passive form of listening. I believe that this robs listeners of the key pleasures of engaging with the artform.
Traditionally, the end of a piece of music, whether that be a record or a playlist, pushes a listener to be proactive with their listening. They have the choice of what to listen to next, including something different to what they had just listened to — or nothing at all. But in the case of autoplay algorithms, tracks come one after another like products delivered from an endless conveyor belt. This discourages listeners from diving deeper into music, reflecting on the songs they have heard and pushing themselves outside of their comfort zone. Instead of sitting with a song or album after it concludes and thinking about what emotions it evoked, the listener is encouraged by streaming services to go on to a similar sounding tune. It turns music into unidentifiable sonic mush incapable of provoking any thought or emotion in its audience. Autoplay algorithms encourage users not to take advantage of their ability to choose. This pushes them away from the opportunity to learn about musical perspectives and ideas which are unfamiliar to them.
One could make the argument that listening to playlists, like Spotify’s “Discover Weekly,” requires a greater degree of listener agency than autoplay. Autoplay is a default feature on most streaming services, like Spotify and Apple Music, and is turned on unless deactivated in the settings. The music continues without you exerting any effort. In contrast, algorithm-driven playlists require users to choose to engage with their music recommendations by selecting them just as they would any other album in a library. If that problem of agency is solved, and playlists like “Discover Weekly” claim to show music outside of a listener’s comfort zone, what is so bad about using them? To me, the problem with these algorithms is that their presentation of new music is more limited than one might assume.
When I was in high school, I had made it a habit of listening to my “Discover Weekly” on Spotify. Each week, I was presented with 30 songs I had never heard before. However, there was still a degree of familiarity to them. What I was “discovering” reflected my existing listening habits and biases. The tracks recommended to me were still predominantly Western, English-speaking, post-1980 and within genres I was familiar with, such as hip hop and rock. With all these qualifiers in mind, the presentation of 30 new unfamiliar discoveries by “Discovery Weekly” is laughable. Its discovery potential is more like a slightly expanded autoplay algorithm, maintaining the “vibe” of a user’s listening habits. The limited scope is a part of the product; their curated algorithms designed to fit your “taste” are seen as a feature, used to attract paid subscribers away from their competition. Presenting subscribers with music that challenges them might disrupt their enjoyment of the service, which is bad business. This includes music from underrepresented cultures, or from outside of a user’s culture at all. That is the distinction between playlists made by these algorithms and collections of music from other humans. Other people have different life experiences and tastes and will challenge you in ways that an algorithm based on your own listening history cannot.
When I think about how I found out about the music I listen to today, these works have entered my life in seemingly infinite ways. A friend told me I reminded them of the title track to Oneohtrix Point Never’s “Replica.” From there, I became curious about music that Daniel Lopatin made separately from OPN. His album “Eccojams” has soundtracked most of my freshman year at Lewis & Clark. I listened to Broken Social Scene’s “You Forgot It In People” because a song from it appeared in a favorite film of mine. I found a review for James Carr’s “At the Dark End of the Street” on Rate Your Music that I found interesting, and it is one of my favorite love songs of all time now. Carly Rae Jepsen’s “Call Me Maybe” was everywhere on the radio when I was growing up, and now I have a hard time not listening to it at least once when I see her album “Kiss” in my library.
In a world where the avenues for discovering music are infinite, why listen passively to an infinite stream of familiarity promoted by an algorithm? Instead, be an active listener, talk to your friends, read about music and put yourself out into a more diverse musical world.
Author
About the Author
