XPost: talk.politics.misc, alt.survival, alt.politics
XPost: alt.science

https://techxplore.com/news/2023-01-microsoft-vall-e-faithfully-voice.html

A team of researchers at Microsoft has demonstrated a new AI
system that is capable of mimicking a person's voice after
training with a recording just three seconds long. The team
explains developing the new app in a paper published on the
arXiv preprint server. They have also posted a webpage
demonstrating the app's capabilities.

Artificial intelligence applications require training on
massive amounts of data. But in this new endeavor, the team
at Microsoft has shown that does not always have to be the
case.

The new app was built using Meta's EnCodec audio compression
technology, and was originally intended as a way to improve
the quality of phone conversations. Subsequent work showed
that it is capable of far more—not only can it mimic a voice,
it can also simulate tone and even the acoustics of the
environment in which the original recording was made.

. . .

Their bragging link is :

https://valle-demo.github.io/

So .... NOW they can combine deepfake video with
deepfake audio - just a few secs of conversation
with your significant reps and they can fake the
voice - even the 'ambience' of their office - so
perfectly YOU can't tell the diff.

Do these tekkies REALLY think they're doing
something GOOD ???????????????

If they have 3-digit IQs then they OUGHTTA very
acutely realize the massive disinfo/crime potential
of their invention. So, arrest them. It's like
inventing a much cheaper way to make fentanyl
or something.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)