(Subscribe to our Today's Cache newsletter for a quick snapshot of top 5 tech stories. Click here to subscribe for free.)
Algorithms help users search, sort and filter music from range of collections. But the suggestions may be inaccurate for fans of non-mainstream music like pop and hard rock, according to a study titled ‘Support the underground: characteristics of beyond-mainstream music listeners’ by the Graz University of Technology in Austria.
The team of researchers used a dataset containing listening histories of 4,148 users of the music streaming platform Last.fm who listened mostly to non-mainstream music or mostly mainstream music.
Also Read | Spotify launches Loud & Clear to shed light on artistes’ revenues and streaming stats
They noticed the algorithm categorised music listeners into four main types: listeners of genres with only acoustic instruments like folk, users of high-energy music like hard rock and pop, users of music with no acoustics and no vocals like ambient music, and listeners of high-energy music with no vocals like electronica.
By comparing each group’s listening histories, they noted that the willingness of a user to listen to music outside of their primary preferences also had a positive effect on the quality of recommendations.
This indicated a bias in music recommendation algorithm, with listeners of high-energy music receiving the least accurate music recommendations and those who mainly listened to ambient music.
Also Read | First-ever non-fungible token-based rock album is out
This isn’t the first time algorithm-based music recommendations have been criticised of bias. A separate study by the Pompeu Fabra University in Spain said that the algorithm is more likely to pick up music by male artists than female artists.
The music streaming industry has become popular today, having more than 400 million subscribers worldwide, according to analytics firm Statista.
