How does Spotify use your data? Even experts aren’t sure
Spotify has revolutionised the music industry, and its ability to recommend music tailored to your personal taste has been a standout feature.
But it isn’t the only app to provide this kind of personalised experience, with Artificial Intelligence being used to create your personalised newsfeeds on Facebook and Twitter, recommend purchases on Amazon, or even the order of search results on Google.
To achieve this, these apps and websites use our data in their recommendation algorithms - but they are so secretive about these algorithms that we don’t fully know how they work.
In a search for answers, a team of New Zealand legal and music experts pored over several versions of the privacy policies and Terms of Use used by Spotify and Tinder to determine how our data is being used as new features have been rolled out.
Their work, published in the Journal of the Royal Society of New Zealand, found that Spotify’s privacy policy has nearly doubled since its launch in 2012, which reflects an increase in the amount of data the platform now collects.
The algorithm hungers for data
Originally, Spotify collected basic information such as the kinds of songs played, the playlists created, and the email address, age, gender, and location of a user, as well as their profile picture, and the pictures and names of their Facebook friends if their profile was linked.
In the 2021 policy, Spotify collects voice data, users’ photos, and location data - and the team of experts have connected this expansion to the patents the company owns.
That same year, “Spotify was granted a patent that allows the company to promote ‘personalised content’ based on the ‘personality traits’ it detects from voice data and background noise,” the authors wrote, suggesting the algorithm has changed to capture voice data.
As for its Terms of Use, the authors found both Spotify and Tinder used ambiguous wording and vague language, despite expectations that it would be somewhat transparent because it is a legal agreement between the platform and its users.
They noted that the opaque style of the Terms of Use made analysis more difficult.
Despite this, they found that from 2015, Spotify’s recommendations were also influenced by “commercial considerations”, including third-party agreements Spotify had with other companies.
The team of experts argue that this particular change “provides ample room for the company to legally highlight content to a specific user based on a commercial agreement”.
Meanwhile, Spotify has also started offering artists the option to lower their royalty rate “in exchange for an increased number of recommendations”.
Taken together, the authors argue that this means that the playlists made specifically for us could be influenced by factors outside of our control, “like commercial deals with artists and labels”.
Users deserve answers
Though they made these findings, the authors note that some will still be speculative while companies stay tight-lipped about how their algorithms work.
“When companies are uncooperative, and typical academic inquiry cannot be complete without breaching contractual agreements, we maintain that scholarly investigation can have a speculative character,” they wrote.
“This suggestion does not mean that a less academic rigour can be expected or granted about making assumptions on the basis of partial, observable data. Instead, we propose that it is the companies’ remit and burden to refute such assumptions and communicating the clarity of their systems.”
With many of us using services like Spotify, Tinder, Google and Amazon on a daily basis, it’s up to these companies to become more transparent in how they use our information with the understanding that we deserve to know what happens to the data that makes us, us.
Image: Getty Images