Until now, Artificial Intelligence has given impressive samples of its possibilities. That is to say, we have seen it from creating images of cats and people that do not exist, to writing emails or designing sports shoes. The so-called Deep Fakes have also been enhanced by this technology and in that sense, there is another similar possibility. At the moment, services capable of cloning voices using Artificial Intelligence are being sold.
Although the mechanisms used to achieve this may differ a bit from each other, the end result is a copy of our voice from a previous listening.
Artificial Intelligence and its potential to clone voices
The first example that we are going to mention on this topic is not yet in its stable version and is in full training. However, it is impressive what Artificial Intelligence is capable of doing. Tacotron to clone voices. This project allows you to replicate any voice by listening to it 5 times, managing to imitate from the accent to the tone you use.
Perhaps the most accessible and complete service of this type is the one offered by Resemble. This company has been in charge of developing voice synthesizers for Call Centers and even games. One of its tools is called Voice Cloning and to use it, we will first have to insert some examples of our voice. From this input, Resemble is able to imitate her and read texts with her.
We can also find respect, a company whose services in the area of cloning voices with Artificial Intelligence have been used by companies such as LucasFilm and media such as Telemundo. The most striking thing about their results is that they promise completely natural voices, imitated or generated from scratch.
In that sense, we have a very useful technology for the media, video games, movies and even advertising. However, this also opens the door to a new generation of much more real and fine-tuned Deep Fakes.