Music radio persists in an era of hyper-personalised playlists, possibly because it offers a contrasting linear, set-and-forget experience.
One of Germany’s most popular radio stations, Südwestrundfunk (SWR), has launched a new radio app with the usual live-rewind and listen again functions, but is also trying something slightly different: AI-powered, in-show song personalisation. (Article in German)
If the user doesn’t like the song the DJ is playing, the new SWR app allows the user to hit “shuffle” – and the app, using AI technology from Berlin startup Cyanite, figures out what the listener likes, then picks another song from the station’s 1300-song “music pool” – and plays that instead.
Cyanite’s AI technology is being used by other German broadcasters. Mediengruppe RTL, which runs a number of large national TV stations, uses their tagging, search and recommendation engine to navigate their in-house catalogue of licensed music.
SWR’s shuffle function, devised by SWR Labs, poses some existential questions about the nature of music radio – if the DJ isn’t talking around songs they have picked, is it still radio? But listeners, and other stations, may approve of this compromise situation: because alternative songs are only drawn from the radio’s wider playlist, the station retains its musical character, even if each listener listens to something slightly different.
To develop this semi-interactive radio playlist, some long negotiations with labels and German music rights society GEMA were needed, but this kind of effort may be deemed worthwhile if listeners prefer their favourite radio show when it only plays music they like.