Spotify has announced a partnership with OpenAI to use the company’s technology in the podcasts distributed by the platform. The artificial intelligence will clone the voices of presenters and participants and use them to “redub” programs in other languages. Thus, audiences who do not master the original language will be able to listen.
For now, the tools have only been used for translation from English to Spanish, and only in a few more famous podcasts. Spotify plans to add French and German translations in the coming weeks.
The translated episodes are already on the platform, in a space dedicated to them, in case you want to check the result.
The feature is based on OpenAI’s Whisper technology, which is able to transcribe speech and translate other languages into English. She is the same one employed in the ChatGPT app to understand voice commands and in WhatsApp bots to transcribe audios.
Spotify’s tool goes further: it creates a synthesized version of the podcasts’ hosts and guests’ voices, as if they were speaking in another language, without requiring them to learn the language and re-record the show.
It’s not the first time Spotify has experimented with using artificial intelligence in its app. The company created DJ X, an AI that functions as a radio announcer, talking to the user between one and another recommended song. The tool is in beta phase and has not reached Brazil.
ChatGPT now also speaks
Spotify did not give more technical details, but it is likely that the voice processing news will also use OpenAI technologies as a base.
The company announced on the morning of Monday (25) a major update to the ChatGPT mobile app, including voice responses. For now, these tools are only available to users of the Plus and Enterprise plans.
