Imagine being able to instantly translate a podcast into another language while preserving the podcaster's own voice.
Spotify is turning this concept into reality with its new AI-powered voice translation feature. The feature utilizes artificial intelligence to reproduce podcasts in other languages, using the original podcaster's voice.
To launch this innovative tool, Spotify has partnered with a select group of podcasters to translate their English-language episodes into Spanish.
The company plans to extend this feature to include French and German translations in the coming weeks. Among the notable podcasters in the initial lineup are Dax Shepard, Monica Padman, Lex Fridman, Bill Simmons, and Steven Bartlett.
Spotify intends to expand the roster to include podcasts like "The Rewatchables" from The Ringer and an upcoming show by Trevor Noah.
At the core of this translation feature is OpenAI's voice transcription tool called Whisper, which not only transcribes English speech but also translates other languages into English.
What sets Spotify's tool apart is its ability to not only translate speech to text but also to reproduce the podcast in a synthesized version of the podcaster's own voice.
Ziad Sultan, Spotify's vice president of personalization, explained, "By matching the creator’s own voice, Voice Translation gives listeners around the world the power to discover and be inspired by new podcasters in a more authentic way than ever before."
While the specifics of the voice replication technology were not detailed, it's likely that OpenAI played a role in this aspect as well.
OpenAI, in a separate announcement, unveiled a tool capable of creating "human-like audio from just text and a few seconds of sample speech."
Spotify, for now, is testing the translation tech with a limited group of podcasters, with broader availability and expansion plans yet to be disclosed.