. |
The A.I can make deep fakes and video-tape manipulation, which deceive the lie detectors.
For artificial intelligence is easy to fool the artificial intelligence-based lie detectors. And the reason for that is that artificial intelligence can "clean" the use of voice and distance of words. Also if that kind of fake-making tool can manipulate the image records or film, it can cut the unwanted movements off the video. The thing that this kind of system requires is just to know what kind of things must remove from the voice and video. And that kind of deep-fake system can also be used in the so-called streaming videos.
Making the stream, where all lies are removed is quite easy. The video and audio are first stored on the hard disk, where the computer or artificial intelligence algorithm makes the manipulation. The people, who are seeing the stream don't recognize that the stream would come one or two seconds later. So the system just parks the data to the hard disk for the second, and the artificial intelligence, which is operating on the powerful computers can just make the deep fakes in a couple of seconds.
And that thing means that this kind of thing is extremely dangerous. The system must just cut the unwanted movements away, and then adjust the voice of the speaker to neutral. The time, what this thing takes depends on how good cuts the system uses. If the system needs to delete things from the single images from the tape, it will take a little bit longer than just removing the part of the tape.
Making deep-fakes is possible also by using the image manipulation programs. In that case, the person can manipulate film from frame to frame, and then release the product to the internet. And in that kind of case, the image-manipulator can put themself to talk with the great statesmen and historical characters.
https://scitechdaily.com/computer-scientists-create-fake-videos-that-fool-state-of-the-art-deepfake-detectors/
Comments
Post a Comment