These platforms want to expand the audience of their content at a lower cost.
In the coming days, French users of Spotify will be able to listen in their language to a two-and-a-half hour interview, originally recorded in English, between American podcaster Lex Fridman and best-selling Israeli author Yuval Noah Harari (Sapiens). But, against all expectations, no dubbing actor is involved in this operation. It is in fact an artificial intelligence (AI) which took care of this task, with voices imitating the tone and speech of the two men. The podcast was automatically dubbed into French, Spanish and German.
This tour de force comes from a new tool from Spotify, called “Voice Translation”, created in collaboration with OpenAI. The creator of ChatGPT is now capable of generating a voice imitation from an audio recording of a few seconds, and making it say any text.
Spotify will test this AI dubbing with six American podcasters, including Dax Shepard and Monica Padman. “ By reproducing the voice of the creator, Voice Translation allows listeners around the world to discover new programs in a more natural way than ever before », promises the company. And this for a much lower cost than dubbing carried out in the studio by professional actors.