Around 2014, I downloaded Spotify for the first time. Since then, I have discovered a bounty of amazing new artists. I have added their songs to my playlists, bought their albums on vinyl, gone to their shows, and worn their overpriced merch.
But I wouldn’t have shaken hands with Japanese Breakfast or seen her show twice if it wasn’t for Spotify’s algorithm. Of course, algorithms aren’t perfect — I get recommended weird stuff all the time — but when I saw Martin Scorsese’s criticism of them in Harper’s, I wanted to defend the damn robots.
To me, the debate on algorithms is less of a conversation on Robots Taking Over The World. Rather, it’s an age-old conversation on gatekeeping.
In this article, I’ll take a gander at the following:
- Martin Scorsese and the Rise of Content
- Algorithms: Pros
- Algorithms: Cons
- Are Algorithms Good for Artists?
- Are Algorithms Good for Artistic Innovation?
- What Is [Good] Taste?
- Monoculture vs. Niche-ifying
- Personal Thoughts
- Further Reading
By the way, if you’re interested in this topic and want to contribute to a potential update article in the future, please participate in my survey!
Martin Scorsese and the Rise of Content
In an essay for Harper’s Magazine, Martin Scorsese — the Academy Award-winning director who made headlines for not counting Marvel movies as “cinema” — decries the rise of algorithms. Algorithms have cheapened film, he argues. Cinema has been reduced to the money-making, eyeball-grabbing business of “content.”
Streaming, Scorsese writes, “has created a situation in which everything is presented to the viewer on a level playing field, which sounds democratic but isn’t.” He wants more human curation, which he believes “isn’t undemocratic or ‘elitist’” because “you’re sharing what you love and what has inspired you.”
But it’s very easy for me to argue this point. Because curation is elitist. And kind of patronizing (Scorsese calls curation “an act of generosity”). How can Scorsese, a 78-year-old white American male, possibly know what I, a 27-year-old Chinese Canadian queer female, would like to watch?
To me, the debate on algorithms is less of a conversation on Robots Taking Over The World. Rather, it’s an age-old conversation on gatekeeping.
Scorsese spends most of his essay nostalgically waxing praises about classic art films and filmmakers. These are works and people I have never heard of, perhaps because I’m not as educated as the great Martin Scorsese.
Not that I don’t appreciate art. I’m a record-collecting, classics-reading, play-going snob who’s been called “hipster” more times than they can count. I also believe people shouldn’t pirate things for free when they’re more than willing to pay for a cup of coffee.
In other words, I believe in art’s value and I do agree with Scorsese’s point that framing everything as “content” cheapens it. After all, that’s a word I use to describe the contents of my underwear drawer.
And things have gotten cheaper, in a sense. The Netflix Era favours more contained series like Queen’s Gambit and Tiger King that are short and sweet. Gone are the days Star Trek and Buffy, shows with twenty-something episodes per season and character arcs that develop slowly over half a dozen seasons.
But is this really the fault of algorithms? And are shorter, catchier shows really The End Of Art As We Know It?
Let’s take a look at what algorithms do well.
Firstly, there is so much media content these days that some form of curation, whether human or robotic, has become necessary. People with families to care for and work to get done don’t have time to sift through The New Yorker and make thoroughly informed decisions on what constitutes quality entertainment.
But the biggest win of algorithms is their ability to personalize the user experience. When I open up Netflix, it feels like Netflix is working for me, not production companies that want to sell things to me.
How can Scorsese, a 78-year-old white American male, possibly know what I, a 27-year-old Chinese Canadian queer female, would like to watch?
Of course, there may be business decisions lurking under the surface that I’m not aware of, that put Show X on my front page and not Show Y. But for the most part, I appreciate that if I watch several K-dramas, I’ll get more K-dramas recommended to me and not horror movies.
And while I don’t have hard evidence in the form of a scientific study to prove this, I strongly believe algorithms are what led me to discover many of the artists I love today.
I like dreamy, shoegazey indie rock with female vocals, so Spotify feeds me Japanese Breakfast, Soccer Mommy, and Clairo. If Spotify didn’t exist and all I had was the radio, I may have never found these artists because I’d have to sit through hours of pre-chosen radio tracks to stumble upon them.
In other words, I like algorithms because they do the opposite of what Scorsese-style curation does. Algorithms feel like it isn’t some network executive or self-proclaimed tastemaker who tells me what to watch or listen to; rather, it’s me and my consumption habits that determine what art gets put in front of me.
Obviously, an individual human’s taste can’t be distilled into a formula. I like dreamy, shoegazey indie rock with female vocals, but occasionally I do like a husky male voice or an aggressive, bass-heavy hip-hop track.
So, I wonder if algorithms that keep feeding me similar things will make me too comfortable in my personal echo chamber to discover new art.
But I think this is easily remedied. Perhaps we can add a “try something new” widget on Netflix and Spotify that recommends wildly different things, Tinder-style, and you swipe right or left on what you like and dislike.
Overall, I’m more concerned with how an algorithm is constructed over whether algorithms should exist or not.
Algorithms feel like it isn’t some network executive or self-proclaimed tastemaker who tells me what to watch or listen to; rather, it’s me and my consumption habits that determine what art gets put in front of me.
According to one guide for artists, certain stats affect an artist’s success with the Spotify algorithm. These stats include skip rate, listening time (apparently, the key is to get someone to listen to your track past the first 30 seconds), playlist inclusion, and even release timing.
Using stats like these to determine whether a song gets exposure or not is problematic, or at least flawed. Just because a song doesn’t hook you in within the first 30 seconds, doesn’t mean it’s a bad song. Besides, popularity does not equal quality, and quality does not necessarily lead to popularity.
Are Algorithms Good for Artists?
Most of us are content consumers, but on the other side of the equation, let’s look at whether algorithms benefit the people making the art themselves. After all, people who make art for a living, such as musicians, are notoriously underpaid and undervalued.
On a platform like Spotify where more algorithm-friendly songs get more airtime, artists are incentivized to make music not for people, but for algorithms. This is obviously problematic.
But if an algorithm competently understands genre similarities, it can help artists find new audiences. If you make music that sounds like Debussy, algorithms can get your music in front of people who appreciate Debussy-sounding music.
I wonder if algorithms that keep feeding me similar things will make me too comfortable in my personal echo chamber to discover new art.
Slate explains the nuances of streaming culture as “both salt and salve.” The problems with streaming are just new iterations of established problems within the music industry, argues Brandon Tensley. These are the kinds of problems that make a small elite rich while the rest languish in obscurity. Because you need a distributor to upload your music to Spotify, it favours established artists who already have the backing power of labels.
On the other hand, Spotify is experimenting with a new model where anyone can upload music, no distributor required. In 2020, Spotify also introduced a new tool that allows artists to have more input on how their music gets recommended.
So, in the realm of music at least, I think the jury’s still out on whether algorithms benefit or take away.
Are Algorithms Good for Artistic Innovation?
Perhaps the place algorithms have had the most negative impact is artistic innovation. As I mentioned before, algorithms incentivize artists to create work that pleases algorithms, not humans.
This is what Scorsese was lamenting, and to be fair, he does have a point. The rise of Netflix and Spotify has coincided with the rise of ambient TV and streambait.
“Ambient TV” (coined by Kyle Chayka) is a term given to the kind of content you leave playing on the TV while you chop vegetables or fold laundry. It’s easy-going, low-effort, low-brow stuff; if you miss a line of dialogue, no biggie. You probably weren’t that invested in the show anyway.
Streambait is similar — these are your “chillhop” and “beats to study to” songs. Songs for background ambience. Songs whose artists you may never see the face of or remember the name to.
We may not be emotionally invested in this stuff, yet it’s very, very popular. So popular that ambient shows made it to the top 10 on Netflix and evidently shocked people.
What Is [Good] Taste?
It’s not that algorithms have ruined our sense of good taste. Rather, they’ve revealed our taste. Unencumbered by TV schedules and bolstered by algorithms that reinforce what we already like, we’re boring people who watch Emily in Paris while we vacuum.
But there’s nothing inherently wrong with ambient TV and streambait. People obviously want to watch and listen to that stuff if they’re that popular, and they have the right to. Put another way: I may not agree with your shallow shows, but I will defend to the death your right to watch them!
The rise of Netflix and Spotify has coincided with the rise of ambient TV and streambait.
It’s tempting to label people who listen to streambait and watch mindless ambient TV as uncultured, lazy content consumers, but is this fair? Partaking in mindless entertainment does not preclude you from consuming more higher-brow content.
I like to put on Archer or Friends when I’m cooking and cleaning, but once my meal is ready, I sit down with an hour of Star Trek: The Next Generation. Netflix’s data probably sees me watching inordinate amounts of sitcoms.
But liking Friends does not mean I must not like Star Trek. Rather, I have different shows for different purposes.
Monoculture vs. Niche-ifying
But if people are truly valuing art less, favouring ambient shows over prestige television, can we encourage everyday consumers to be “better”?
We can’t force people to read books and go to symphony orchestras. That’s classist and patronizing, plain and simple. But perhaps we can find novel ways to engage audiences and tell meaningful stories at the same time.
This involves getting to know audiences on a deeper level, and I think personalizing — even if it involves algorithms — is the way to go.
A parallel conversation to the rise of algorithms is the decline of monoculture and the growth of ever-more specific niches. As our movie-watching and music-listening experiences get more personalized thanks to algorithms, entertainment becomes niche-r.
Partaking in mindless entertainment does not preclude you from consuming more higher-brow content.
While some people lament the fall of monoculture — no longer will the world gather round to watch Game of Thrones together — I don’t see niche-ifying as a bad thing. After all, people are different. I, for one, never got into Game of Thrones because I had trouble telling apart all the different bearded white dudes.
So, I never experienced the mass cultural bonding that is waiting for the next episode of GoT to drop. Instead, I get my sense of community from smaller, niche-r shows that resonate with me on a more personal level.
Nichification is also reason to believe that we’re not entirely beholden to algorithms. People still watch things because real humans tell them to.
Case in point, my girlfriend and I watched Teenage Bounty Hunters because it was recommended by Rose and Rosie, my girlfriend’ favourite queer YouTubers. The show’s fanbase isn’t big and it was cancelled after just one season. Yet it has a small, dedicated fanbase who appreciate its female queer representation.
As a young Millennial, I remember the last gasp of cable television before the streaming giants took over. Back then, it was easier to bond with people over shows because we all accessed the same handful of channels and watched the same handful of programs in that hour after school.
But my bonds with others who grew up on Suite Life of Zack and Cody and Teen Titans are not the same kinds of bonds I have with fellow lovers of Wynonna Earp and The Wilds. Cable TV bonding was more regional, but less specific and more superficial.
Today, I can bond with people all over the world. And this bonding is a lot more meaningful — we connect over very specific shows that resonate deeply with the communities we belong to.
And yes, my attention span has decreased. I have been spoiled by ten-episode seasons binge-watched over a weekend, and I can’t imagine having to wait an entire week for the next episode again.
Yet I see these trends as opportunities, not drawbacks.
I get my sense of community from smaller, niche-r shows that resonate with me on a more personal level.
Shorter shows can be better, as The Ringer points out. Back in the day of cable TV, shows were forced into certain formats. This is no longer a concern with streaming. Writers no longer have to stretch stories out longer than necessary, and any good writer knows that telling a story in fewer words is better than being long-winded.
I also see nothing wrong with binging (other than missed homework assignments and poor exam scores). Binging has allowed me to soak up a show and join its fan community very quickly. I am also more emotionally invested in a show if I binge it than if I wait days between episodes.
And while the audience pool for each show may be smaller, that doesn’t mean entertainment quality needs to get worse. Perhaps instead of big-budget special effects and A-list stars, we can have smaller projects with more focused storylines that feature up-and-comers. This is an opportunity to give more airtime to historically under-represented communities like Black, Indigenous, and other artists of colour.
If we do it right, we can avoid a system where big celebrities with the backing of major record labels and studios win everything. We can build a system with more niches, more diversity, and more individuals who each get a slice of the audience attention-cake.
I don’t know how we can accomplish this — though there have been some good ideas like a government-run Great American Music Library — but I do know that if humans created robots, we can make the robots work for us.
Thanks for making it to the end of my essay! If this topic interests you and you want to contribute to a potential update in the future, consider participating in a short survey.
Further Reading (Works Consulted)
If you’re interested in this topic, I recommend reading the following articles that I used to write mine:
- “Il Maestro” by Martin Scorsese in Harper’s Magazine (2021)
- “Are streaming algorithms really damaging film?” by Alex Taylor for BBC News (2021)
- “‘Emily in Paris’ and the Rise of Ambient TV” by Kyle Chayka for The New Yorker (2020)
- “Can monoculture survive the algorithm?” by Kyle Chayka for Vox (2019)
- “The Success Of Streaming Has Been Great For Some, But Is There A Better Way?” by Paula Mejía for NPR (2019)
- “Spotify May Reinforce Many Music Industry Power Imbalances, but a Few Artists Are Using It to Upend Others” by Brian Tensley in Slate (2019)
Li Charmaine Anne (she/they) is a Canadian author and freelance writer on unceded Coast Salish territories (aka Vancouver, Canada). Her work has appeared in literary journals and magazines and she is at work on her first novel, a contemporary YA about queer Asian skater girls. To read Charmaine’s articles for free (no subscription required), sign up for her newsletter.