Menu Menu

Spotify’s speech tech to recommend music based on your mood

Spotify could soon recommend new songs and artists based on analysis of its users voice data. Is there reason to be concerned for our privacy?

Whether you’re feeling glum, upbeat, pensive, or perhaps just a little bored, Spotify will soon recommend music to match your exact mood. Freaky eh?

After a two year wait, the music streaming giant has just been granted a patent for new technology that will soon analyse users’ voice data – including both speech recognition and background noise.

Taking into account audio cues from the environment, a person’s age or gender, accent, and emotional state, the platform is striving to tailor-make music and podcast suggestions for every subscriber in every situation. Current metrics based on listening habits and preferred genres just aren’t quite invasive enough, apparently.

Originally filed in 2018, the 11 page patent outlines Spotify’s method for ‘processing a provided audio signal,’ and then following up by ‘identifying playable content based on the processed audio signal content.’

It sounds like a load of jargon I know, but essentially this means that Spotify will pull live audio data from your smartphone, from which it will determine your current emotional state based on ‘intonation and units of speech’ and also your location type based on background cues, such as bird noises or traffic.

If you’re feeling melancholic on a lone walk in the park, theoretically, you can expect some Daughter or Foals to crop up in suggested artists, and if you’re at a party with friends looking to grab the aux, presumably chart topping bangers from the likes of Justin Bieber and Labrinth.

In any case, this new update arrives as part of Spotify’s attempt to better target its existing audience and those it’s hoping to bring on-board. Already known for algorithms that constantly suggest new content, Spotify believes giving people more of what they want to hear is the key to growing its subscriber base.

Not to mention, it also allows for targeted ads to reach those using the app’s free version.

Of course the topic of privacy vs personalisation comes into question here, and in a pretty big way. In recent years we’ve seen the impact that the unsolicited recording of public data can have, what with Facebook’s Cambridge Analytica scandal, and the miss-use of data from fitness tracking companies and adult streaming sites.

However, in the same breath, we have become accustomed to comprehensive voice assistant technologies capable of running our entire households. With the pandemic keeping people locked up for months at a time, you could argue that Alexa is likely in the 10 most frequently spoken names of 2020.

By and large, we are accepting that a certain level of intrusion is part and parcel of having the most convenient technologies available – but the chief factor is that our consent must be given.

Spotify listeners may very well perk up at the prospect of having a super personalised experience on the app soon, but the methods for collecting our data will have to be outlined in full, and not disguised in the small print of the terms and conditions.

Accessibility