Xavier “X” Jernigan, Spotify’s Head of Cultural Partnerships, has achieved something remarkable: he has become the first voice model for the music streaming platform’s AI-driven feature, their AI DJ. The AI-powered feature offers personalized music based on individual preferences, as well as providing commentary between songs in the same way a real radio host might.
In order to capture Jernigan’s tailored voice, the team at Spotify recorded the podcast host in a studio, getting each inflection, cadence, and word pronunciation exact. They worked with both Sonantic, an AI startup Spotify acquired, as well as OpenAI technologies. Through this process, they were able to create an AI voice that was as realistic as they could make it – passing the ultimate test, Jernigan’s mother.
The AI’s visuals are also designed to be approachable and make the experience more personal. The green circle that the users see is based off of the Spotify logo and its motion mimics the motion of a mouth when the AI speaks.
Jernigan was not just a voice model for the AI, he was also part of the writer’s room for the audio guide. He was able to offer expert commentary of the music, as well as tweak the delivery of phrases that he uses in everyday language.
At the time of release, DJ was only available in English for U.S. and Canadian Premium members, however, in the future the streaming platform might introduce other languages.
Spotify has become one of the biggest music streaming services in the world and is constantly innovating their platform and providing better experiences for their users. With the introduction of their AI DJ, they provide a unique way for people to get personalized recommendations tailored to their interests and preferences – with a human-sounding voice.
