In the near future, any human will be able to clone sound with the help of artificial intelligence. For this, you will need to read the text for about 30 minutes with the microphone. Then when you type any text, it will "read" it with your voice, the Verge writes.
Specialists say that the synthesis of speech has been developing rapidly in recent years. In the past, the realistic composition could be achieved by extracting parts of phrases from audio recordings, but now neural networks generate human speech from scratch. The result sounds faster, easier, and more convincing.
So in the near future, lots of cloned voices of celebrities will appear, which will allow them to be licensed. Such a sound, for example, can be used in advertising so that its owner does not even have to go to a recording studio.
Experts say that artificial intelligence may not be able to convey accurate and versatile intonation, although this is not necessary for automated applications and voice aids.
However, there are risks. If such a voice is caught by a swindler, he may call citizens, ask for money through the voice of a relative or friend, or commit some other machinations.
