Tech

Artists Are Telling Spotify To Never Use 'Emotion Recognition'

The company has a patent for a system that recommends music by monitoring emotional cues in your private conversations.
GettyImages-1232387222

Hundreds of recording artists, human rights groups, and academics have penned an open letter to Spotify, requesting that the company does not use a recently patented technology that would listen to users’ conversations and then recommend content based on their perceived emotions.

The Swedish company began working on the new system in at least 2018, when it applied for a patent, which was approved this January and first reported by the BBC. In an April letter to the digital rights group Access Now, Spotify said it has never implemented the emotion recognition recommendation engine. 

The decision to patent the technology was “influenced by a number of other considerations, including our responsibilities to our users and to society at large,” the company wrote. It declined to comment further for this article.

The coalition of artists and human rights advocates said they appreciated that Spotify was not currently using the technology, but questioned why the company would even explore emotion recognition–which has been shown to be flawed on numerous occasions. The kind of system described in Spotify’s patent would likely lead to manipulation without user consent, discrimination, and harmful privacy breaches, according to the letter.

“Monitoring emotional state, and making recommendations based on it, puts the entity that deploys the tech in a dangerous position of power in relation to a user,” the group wrote. 

It also raised concerns that “using artificial intelligence and surveillance to recommend music will only serve to exacerbate existing disparities in the music industry. Music should be made for human connection, not to please a profit-maximizing algorithm.”