Weird Music #8: Algorithms we live by
For those of us who find it difficult to do go about our everyday tasks without the assistance of a computer, it’s a constant struggle to maintain a safe distance from technology and convince ourselves we could operate without it. Recently, however, I have had my doubts — especially because of my love of music.
Many music aficionados hold that their refined tastes and extensive knowledge of music are entirely their own work, but how can this claim hold water in the digital age? Can one really live a fruitful and wholesome life as a music-lover without a heavy reliance on technology?
The answer, I think, is a cold, resounding “No.”
In days (quite far) past, where a love of music meant being involved in a community, going to shows, and interacting with musicians, there may have been some room for argument; but in the present day, the point seems almost moot. Listening to music depends on what equipment you have. Storing music depends on what device you have. And discovering, acquiring and sharing music with others all depends upon which of the many online music services you deem fit to use.
When it comes to these online music sites, I want to draw particular attention to Spotify — a service that seemed innocent and neutral enough, but one that houses many hard truths for any proud music lover. At first look, Spotify seems to be merely a database of music, an immeasurable library of known and unknown music ripe for the searching. But upon closer scrutiny, it seems the user experience is overwhelmingly driven by user data collection and recommendation algorithms.
Spotify utilizes what is called “Collaborative Filtering,” a complicated process explained by Spotify Engineer Erik Bernhardsson that amounts to little more than correlating separate user ratings and listening habits, and structuring the user’s experiences based on this data. This includes not only the music you are recommended, but also the content of the Spotify radio stations, as well as the specific content of advertisements you are played.
Many other music services use this type of algorithm. Pandora, Last.FM, iTunes and other services are all based upon the idea of “Synthetic Personalization,“a concept developed by linguist Norman Fairclough, whereby large corporations rope in individuals that use their service by providing an “individualized” experience — but in reality one that is mass produced for all users through algorithms.
So what does this mean for the average music lover? It means that one’s personalized experience with these services is merely a very deceptive and effective business model and marketing strategy. It means the formerly glorious process of discovery has been replaced by algorithms feeding you songs you are statistically more likely to appreciate.
I don’t want to sound like some Luddite who’s entirely opposed to using these services — just make sure to leave some time to explore your interests for yourself.