Spotify will clone the voices of podcasts to translate them into other languages

What if you could clone yourself? voice to speak in a language What do you know? This possibility is increasingly real thanks to the artificial intelligence (AI). Aware of this, Spotify announced this Monday the launch of a new function that translates episodes of some of the podcasts best known in the Anglo-Saxon world.

The Swedish audio platform has already made the Spanish versions of the programs available to the public. Lex Friedman, Steven Bartlett, Dax Shepard, Monica Padman either Bill Simmons. In a second round the group will be expanded to translate other popular authors.

Although they may be little known in Spain, these ‘podcasters’ in English have accumulated tens of thousands of followers. Spotify’s pilot program that clones your voice into the third most spoken language in the world, then, will allow them to expand their listener base, thus eliminating linguistic barriers. It is expected to expand to French and German in the coming weeks.

“What if I told you that you could offer your podcast to any listener in the world, in their own local language, but maintaining your own voice? “It’s crazy,” he celebrated. Daniel Eckfounder and CEO of Spotify.

Artificial intelligence

Spotify is the last of the large companies in the digital ecosystem that integrates the Artificial intelligence (AI) At your service. This new feature is possible thanks to a partnership with OpenAIthe emerging firm behind the popular ChatGPT. Thus, the platform will use Whisper, a voice transcription tool, to be able to transcribe the content of these podcasts and then translate it into Spanish.

Related news

Spotify will go a little further and adapt that translated text to audio, using AI to clone the voice of the ‘podcaster’ in question so that it sounds as convincing as possible. “By adjusting to the creator’s own voice, Voice Translation gives listeners around the world the ability to discover and be inspired by new podcasters in a more authentic way than ever,” said Ziad Sultan, vice president of personalization at Spotify, in a statement. a release.

That feature was likely also developed with the help of OpenAI, although it is not official. Just this Monday, the company announced ChatGPT improvements, which will now be able to, among other things, create “human-like audio from text and a few seconds of sample speech.” OpenAI is reducing access to that feature for security and privacy reasons. Perhaps that is why Spotify has also limited it to a small number of podcasters and has not given more details about this new tool.



ttn-24