Google's equivalent in China, called Baidu, has announced very interesting success in the area of artificial intelligence: the Chinese company has created software that can clone a human voice after analyzing it in a few seconds. The software uses state-of-the-art neural network technology and can not just clone the voice, but also change it to reflect another gender, or even sound a foreign accent. The previous attempts at this technology were successful, but only after analyzing much larger "samples" of the votes.
For example, Acad has a software called VoCo that can mimic voice after analyzing 20 minutes of audio. And the program of the Canadian company Lylebird can clone a voice after analyzing just one minute of audio. Chinese software is doing the same after just a few seconds, which is a really impressive achievement. The possibilities are truly unlimited: from cloning the voice of a parent who travels to read his baby's good night's tales to the restoration of historical speeches and documents read with the voices of historical personalities left from just a few seconds of archive audio fragments. Of course, the security aspect must not be forgotten: given that the voice of every person can be cloned , he is no longer as sure as a unique access control method.