Microsoft recently announced their new VALL-E AI text to speech synthesizer that can simulate or mimic a voice using just a 3-second audio sample. Considering how cybercriminals and cheaters are using new AI tech so quickly, my kneejerk reaction to this was a resounding “BUT WHY!?” (cue Jackie Chan meme) but thankfully this might not be the case... yet.
First of all, while the research has been published on github, these are only the results with no code available. Thankfully, Microsoft has decided not to make it as available to try as some other AI tools out there. Secondly, VALL-E is not always spot-on identical all the time. Trying out the results shows that most of the time, the synthesized speech sounds somewhat mechanical and unnatural.
However, there are times when VALL-E does get it right and to my untrained ears, it sounds almost identical. While Microsoft has not made the VALL-E AI available to the public, it’s a definite sign of things to come. Banks in particular are going to have to rethink how they approve loans and other financial decisions via voice calls or we’re all going to stop answering unknown phone numbers in the future (which isn’t good if you’re lost in the woods).
Are you as worried about all these new AI tools as we are? Or do you have a different point of view? Let us know on our Facebook page and stay tuned to TechNave.com
COMMENTS