Would You like a feature Interview?
All Interviews are 100% FREE of Charge
With nearly 100 million songs available and more than 600 million subscribers, helping Spotify find the music they love is a navigational challenge for Spotify. The promise of personalization and meaningful recommendations that give more meaning to our vast catalog is at the heart of Spotify’s mission.
The streaming audio giant’s suite of recommendation tools has grown over the years. Spotify home feed, discover every week, blend, day listand mix made for you. And in recent years, there have been signs that this is working. According to data released by Spotify, Investor Day 2022the number of monthly artist discoveries on Spotify has reached 22 billion, up from 10 billion in 2018, and is “still a long way from reaching that goal,” the company said at the time.
Over the past 10 years, spotify We have invested in AI, particularly machine learning. The recently launched AI DJ may be the company’s biggest bet yet: technology will allow subscribers to better personalize their listening sessions and discover new music. AI DJ mimics the atmosphere of radio by announcing song titles and track introductions. This is meant to make it easier for listeners to step outside of their comfort zone. The existing problem with AI algorithms is that while they can be great at giving listeners what they already know they like, they predict when they want to step out of that comfort zone.
AI DJ is a combination of personalization technology, generative AI, and dynamic AI voices that allows listeners to tap the DJ button when they want to hear something new or something not directly derived from their established tastes. Masu. Behind the sweet sounds of AI DJs are technical and music experts who aim to improve the recommendation capabilities of the Spotify tool. The company has hundreds of music editors and experts around the world. A Spotify spokesperson said generative AI tools allow human experts to “extend their innate knowledge in ways never before possible.” Â
Data about a particular song or artist captures several attributes, such as specific musical characteristics and with which songs or artists it is typically paired. It’s a very simple process to gather information about a song, from year of release, genre, and whether it’s happy to danceable to melancholic, among the millions of listening sessions that the AI algorithm has access to data on. Various musical attributes such as tempo, key, and instrumentation are also identified. Combining this data, tied to millions of listening sessions and other users’ preferences, helps generate new recommendations, allowing for a leap from aggregated data to individual listener assumptions. Masu.
In the simplest formulation, the AI says, “If you liked Y, you also liked Z. We know you like Y, so you might like Z as well.” Here’s how to find out. And Spotify says it’s working. “Since we launched DJ, when DJ listeners hear commentary along with personal music recommendations, Try something new (or listen to a song you might have been skipping),” the spokesperson said.
If you succeed, it’s not just the listeners who will be relieved of their pain. A good discovery tool is equally beneficial for artists looking to connect with new fans.
Julie Knibb, Founder and CEO tomorrow’s music It aims to help artists connect with more listeners by understanding how algorithms work and how to better work with them. Everyone is trying to figure out how to find a balance between familiarity and newness in a meaningful way, and everyone says they’re relying on it. AI algorithms make this possible. She says the balance between discovering new music and staying within established patterns is a central unresolved question for everyone involved, from Spotify to listeners to artists.
“Any AI is only good at what it’s told to do,” Kunibe said. “These recommender systems have been around for over a decade and have gotten very good at predicting what users like. It’s about knowing what’s in the user’s head when you want to step into a category.
Spotify day list uses generative AI to shape and reshape listener preferences throughout the day, not just established preferences, but also different contexts that can create new recommendations to suit different moods, activities, and vibes. This is an attempt to take into consideration. Kunibe said that while these improvements are likely to continue and that AI gets better at finding formulas to calculate how much novelty listeners want, “people are discovering new music all the time.” “The assumption that we want to do that is not true,” he added.
Most people are still quite happy to return to familiar musical territory and listening patterns.
“There are different profiles: listeners, curators, experts, and people make different demands of AI,” Kunibe said. “Professionals are more difficult to wow, but they aren’t the majority of listeners and tend to be more casual.” And their use of Spotify often creates a “comfortable backdrop” to their daily lives. She says it will be.
Technology optimists often speak of the era of “abundance.” With 100 million songs available, many listeners prefer the same 100 songs millions of times, so it’s easy to see why a new balance is needed. But Ben Ratliff, music critic and author of Every Song Ever: Twenty Ways to Listen in an Age of Musical Plenty, says algorithms are more likely to perpetuate the problem than to solve it. I say it’s a thing.
“Spotify is great at capturing the public’s sensibilities and creating soundtracks to match them,” Ratliff said. “the sad girl starter pack For example, the playlist has a great name and about 1.5 million likes. Unfortunately, under the banner of a gift, SSP simplifies the oceanic complexities of youth depression into a small collection of believable “aspirational” musical acts and staunch clichés of music and sensibility. Form more quickly. ”
Curated works that are clearly made by real people with real tastes remain Ratliff’s favorite. Even the best playlists, he says, can be created without much intention or conscience, with a well-developed sense of pattern recognition. “It’s about ambiguities and patterns in things that are widely known,” he said.
AI will vary from person to person, but in a world of 100 million trucks, it is equally likely to become a utopia or a dystopia. Ratliff says most users should keep AI simpler in their streaming music journey. “As long as you understand that the app will never recognize you the way you want it to know you, and as long as you know what you’re looking for, or with some appropriate prompts. You can find lots of great stuff as long as you’re prepared to listen to music on Spotify.