China’s Google Equivalent Can Copy People’s Voices In Just Seconds

One of the biggest threats in the era of digital footprints is the possibility of having one’s identity stolen. Identity theft is one of the most difficult crimes to fight, especially for those who haven’t been guarding their identities as well as they should have. With a person’s voice being part of their identity, therefore, the prospect of it being cloned and used for all sorts of nefarious purposes is quite troubling. This is what China’s Baidu has just made possible.

For those unfamiliar with the company, Baidu is basically the Chinese equivalent of Google, which offers internet search capabilities at a much more restricted level. The firm recently released a white paper detailing a rather alarming new technology. Apparently, engineers were able to develop a technique that can clone voices with only a few samples in just a matter of seconds.

“Voice cloning is a highly desired feature for personalized speech interfaces. Neural network based speech synthesis has been shown to generate high quality speech for a large number of speakers. In this paper, we introduce a neural voice cloning system that takes a few audio samples as input,” the paper’s Abstract reads.

In terms of the result, not only is the technology able to mimic people’s voices with disturbing accuracy, the voices themselves can be changed in different ways. For example, it can be made to sound higher or deeper, to be male or female, or to even have a different accent altogether. There are samples available for the curious on GitHub, which showcase just what this technique can do.

As Futurism notes, there is actually some potential for positive uses of such a technology. Being able to read stories to children in their parents’ voices while the real people are away could be one. However, this could also be used to cause substantial damage, especially when used on politicians.