The platform is heavily oversaturated and for labels wanting their next big money-maker, sorting through all those tracks is an impossible nightmare. Signing a handful of artists in hopes that one will become majorly successful is also a gamble.
The solution? Algorithms.
A new software called Musiio has been developed to run on streaming services, scanning through thousands of songs, and categorising them according to genre. Musiio is also capable of finding similarities between a demo online and an existing top-charting hit.
Other algorithms are trained to process patterns in digital data, where large amounts of online activity can indicate a trending artist. For example, where a song is used on TikTok millions of times, or high engagement on an emerging artist’s YouTube video.
The software is named Sodatone and can also detect mentions on music blogs, high booking numbers for venues, and the number of times a track is included on playlists or charts. Promo from influencers is also significant to help labels find upcoming artists.
While this type of technology has only emerged in the last three years, AI and algorithms have been part of the music listening experience for many more.
Think about your ‘Discover Weekly’ playlist, or the much anticipated ‘Spotify Unwrapped’ that analyses your annual listening habits.
These custom-made playlists gather tons of data on your listening habits to recommend more music that fits your tastes.
The use of algorithms in the streaming industry has been contested often. For example, if artists know a song only needs to be played for 30 seconds to count as a play, then of course they’ll start placing some of their album’s best moments at the start of a track.
In other cases, bias has been highlighted as a problem for music algorithms. As always, whoever codes the algorithm will unwittingly incorporate some of their own bias into it, meaning that racial and gender biases exist within.
A study on recommendation algorithms showed that apps like Spotify are more likely to suggest male artists over female artists. However, these biases can be removed, once correctly identified, and addressed.
When algorithms are left to discover artists who make music synonymous with the current Top 40, are we at risk of hearing nothing but the same thing repeatedly?
Probably not, because this has already been happening in popular music for years.
Many believe that while algorithm use is on the rise within the music industry, they aren’t going to completely takeover just yet.
There are limitations to what AI can do with music. Popular genres such as pop, indie, trap, and UK grime are easily detected, but niche types of cultural music still can’t be well sorted.
When it comes to culture and arts – both very subjective experiences – it sounds like a human touch is necessary to find true talent. At least for now.