Spotify patents tech to recommend songs based on users' speech, emotion
- Bryan Walsh, author of Axios Future
Spotify logo. Photo Illustration: Omar Marques/SOPA Images/SOPA Images/LightRocket via Getty Images
The music-streaming company Spotify was granted a patent for technology that aims to interpret users’ speech and background noise to better curate the music it serves up.
Why it matters: Aside from being a little weird and invasive, the technology is an example of a future trend in computing: emotion recognition.
What's happening: Music Business Worldwide reported this week that Spotify had filed in February 2018 and been granted this month a patent that uses "speech recognition to determine [users'] 'emotional state, gender, age, or accent' — attributes that can then be used to recommend content."
How it works: According to the patent filing, the company is developing technology that could extract "intonation, stress, rhythm, and the likes of units of speech" that would permit the "emotional state of a speaker to be detected and categorized."
- Combined with other data from a user's listening history and past requests, appropriate music could then be recommended or played.
What they're saying: Not surprisingly, the internet had fun with this one.
The catch: Technology companies often file patents for innovations that are never used in their products, and a company spokesperson told Pitchfork, "We don't have any news to share at this time."
What's next: Whether or not this capability ever makes it into Spotify, companies are increasingly exploring technology that purports to recognize emotional states through voice tone.
- Amazon's new Halo fitness tracker analyzes users' vocal tone to evaluate how they're coming off to other people.
- But there are concerns that emotion recognition could be misused — a report released this week from a U.K. human rights group identified dozens of companies in China using the technology, including some working with the police.
The bottom line: I'm a Spotify user, but the company wouldn't need to read my tone to serve up emotionally appropriate music.
- Happy? The National. Sad? The National. A bit phlegmy? The National.