Voice Deepfake Allows You to Create Computer Generated Speech


Deepfake: Recently, Veritone, which specializes in the development of artificial intelligence products, launched Marvel.AI, a tool capable of “cloning voices” – also known as “voice deepfake”. According to the company, brands and companies can use the software to create monetized audios without the need for humans.

Veritone detailed the proposal in an official statement, issued last Friday (14): “with complete control over your voice and your use, any influencer, personality or celebrity can literally be in several places at the same time,” he comments. Initially, as the company explains, the tool should be used to expand the participation of individuals in advertising projects and similar types of media.

Marvel.AI works like a regular deepfake, but uses audio recordings and not images in the Machine Learning process – a user or company can, for example, “teach” how the software should replicate a specific voice. In this way, Veritone customers can request custom clip orders or choose a voice model from a database.

In this context, it is inevitable to say that the advance of this type of technology is impressive. But, on the other hand, its use and business model raise doubts about several ethical issues. In the case of Veritone, the holder of the “rights over the voice” (which, at times, may not necessarily be its “natural owner”) may request any type of personalized audio and imply situations of misuse.

Ryan Steelberg, director of Veritone, commented on the matter in an interview with The Verge website: “Whoever has the copyright to these voices, we will work with them to bring them to market,” he explains, “that will be up to the rights holder and what they think is appropriate, “he concludes.


Please enter your comment!
Please enter your name here