Making a living as a musician, especially an independent artist, is extraordinarily difficult. Add artificial intelligence and its broad influence on the music industry, and it becomes even harder to navigate.
The University of Toronto recently hosted a panel discussion on the topic with a focus on indie artists. The goal is to turn out grads who can function in today’s music environment.
The panel discussions and talks, presented as part of the Toronto Music Entrepreneurship Exchange, more broadly, part of University of Toronto Entrepreneurship Week, and covered the way AI affects everything from algorithm-driven streaming to artistic ownership.
The Initiative
Ely Lyonblum is a Strategic Research Development Officer at UofT’s Faculty of Music, and he set up the panel discussion as part of his role.
“We have that role in each division across the university,” he explains. As a Strategic Research Development Officer, he’s embedded within the music department, where he gets to know the faculty, students, and how to support them and their research in creative ways.
“I have the luxury, the privilege and the joy of really being able to connect with students and educators.” At the same time, he can compare notes and coordinate with his peers in other departments if needed.
Lyonblum completed his own Doctorate in Music a few years ago. He recalls hearing a lot of conversations around the notion of music entrepreneurship. An indie artist has to think like a business person, and understand the environment they work in. As he points out, even for purely academic purposes, understanding the way the industry works is key.
“I wanted to make sure that knowledge was shared as widely as possible.” He says the students learn practical skills. “I see the work that we’re doing as very complementary.” Along with a graduate seminar, Ely routinely connects with professors to help amplify what they want to convey to their students.
AI isn’t a compulsory part of any music program, but as he points out, it enters into everything from music creation through text or music generation, or the way it is used by streaming music services to expose specific tracks through recommendations.
Spotify: How AI Suggests New Music
On streaming services like Spotify, AI models create the recommendations that point users to their next favourite music, podcasts, or other content.
An algorithm makes predictions based on the data it gathers on a user’s behaviour and preferences. That data is compared with other users with similar profiles. It can also base its recommendations on the characteristics of the music itself; i.e. it looks at the specs of a track you like or save, and finds others with similar elements.
Spotify’s algorithms use a hybrid approach that should consider both aspects, i.e. user history and the specific songs streamed.
“There are so many ways they will encounter it,” Lyonblum says. He feels it’s an important part of music industry literacy in the 21st century. “We don’t want there to be a sense of unfounded fear.” There are real concerns about replacing composers with AI, and other issues, he acknowledges. “But, we want to give our students as much knowledge about where those concerns comer from.”
Working With The Algorithms
Daniel Field is an independent musician with an ambient music label called Imaginary North. He records music under the name Kilometre Club, and he currently counts more than 120K monthly listeners on Spotify, with his biggest track as more than 1.8 million streams.
Figuring out how to use the algorithms in his favour led to a big part of his current — and unexpected — success.
“My context is, I’m currently a middle school teacher,” he says. Since he was a child singing folk songs, he’s also worked as a musician on a casual and part-time basis. “I enjoyed the recreation of being a musician.”
The slowdown of the pandemic led to learning about synthesizers, and branching into ambient music. It also led to a sense of real opportunity. “I never really tried in a meaningful way,” he says of his musical ambitions up to that point. As he delved into the genre, though, it changed his mind about what was possible. “There was an obvious, and growing, niche in ambient music,” he says.
Producing and recording music is now possible with a synth and a laptop; finding listeners is the real trick. He says he approached it from the “standpoint of knowing nothing”.
Like any other indie artist, he searched for ways to attract the attention of those Spotify algorithms. “They have an algorithm built in, [so] how do you find yourself in the flow of the water?”
Daniel tried different approaches, and found that the details really matter. From trial and error he says he discovered, for instance, that ambient tracks with drum beats tended to get fewer streams. Add a few bird sounds, however, and the opposite effect seemed to come into play.
If you can find a way, do you simply go with that flow? “Do you need to kind of risk your own artistic integrity to placate the algorithms?” he wondered. From the creator’s standpoint, do you conform to what seems to feed the algorithm what it wants?
As he points out, it raises other issues. The algorithms and playlists tend to group tracks according to their feel or vibe, rather than targeting specific artists.
“You’re still essentially anonymous,” he says. “They’re turning music into vibes.”
Playing only to what works also takes a creative edge out of the music. “It’s kind of softening all music,” he says. “Are you offering a vibe? Or are you offering an experience?” Younger listeners, he mentions, often tend to listen to specific genres — not the artists themselves.
Artists looking for recognition in their own right, rather than simply ears to listen to their tracks, need a completely different approach.
Daniel notes that streaming services aren’t music companies. “They’re in the world of selling advertising,” he says. “It’s also adding to the difficulty.”
When it comes to ambient music production, Field points out, the distinction between AI and composer can be blurred from the outset. “What happens when you use an arpeggiator or chord randomizer — technically, that’s also AI,” he says. “Where is the line drawn? I would argue, even if you went back to the 80s when they first had randomizers and arpeggiators […] do we call that AI?”
AI generated music is a far more complex issue that it may seem.
Success in the world of ambient music has changed his life. These day’s teaching and music are both part-time.
“Were seeing the term AI as the poor man’s […] laziness,” he says, “but even Garage Band has presets.”
Is More Better? And… Who’s In Charge?
Vanese VJ Smith is the co-founder of Loop Sessions Toronto. Part of an international organization, it works to support music production and education, with a focus on vinyl record culture. As Pursuit Grooves, Vanese has performed as a DJ and electronic music creator for two decades, and mentored students through Toronto Metropolitan University, Artscape, Canadian Music Centre, Small World Music, and Disney, among other prominent organizations.
Loop Sessions Toronto organizes monthly events around music culture, production and creativity. “We started in 2019,” Varese says. The program has grown and expanded post-pandemic.
“I am myself a creative.” Smith is a music producer and sound artist, and her opinions are rooted in history. She works in the community as well as academia.
“I’ve been very observant of particular ways in which we — especially in the creative sector — are affected by technology.” Different generations also approach it differently. “I’m also clear that someone half my age, the way they engage with technology is very present.” They can’t compare before and after, in other words.
“I can’t speak as to algorithm as a community, because we all come from different places,” she says.
She talks about her own approach to using and teaching about technology in music creation. “My approach to mentorship and to education is very much geared towards the individual using art as self expression,” she begins. It’s about encouraging authenticity, and ways to maximize limitations. “Maximize what you have to push your creativity until it doesn’t work for you anymore — before you move along to the next thing,” she explains.
Using AI generated options from the outset can present a different sort of problem: too many options. “It’s like going to a restaurant when you have 2,000 things on the menu,” she says. “It seems better, but is it more creative?” she asks.
“I come from a background where I started with very simple gear,” she says. Limitations force more creative thinking.
Working with algorithms for music exposure can be a tricky proposition. “Spotify is one of those strange, strange, strange things, because you’re talking about the consumer vs. the artist,” Vanese points out. It offers consumers endless choice, and artists… fewer and fewer means of making money from it.
“This is an industry that has never supported an active living wage for its workers,” she says. “How we discover things has become very streamlined. All of these questions really depend on what side of the fence you sit on,” she adds.
“I definitely have more questions from the artistic perspective,” Smith says.
“It would be nice if there were more artists in control of these initiatives.”
Are you looking to promote an event? Have a news tip? Need to know the best events happening this weekend? Send us a note.
#LUDWIGVAN
Get the daily arts news straight to your inbox.