AI can detect deepflake videos

[Graybeard]

Intel has devised a way to detect deepflake videos and help prevent the amplification of these videos in social media as well as being further amplified by the news reporting mainstream media.

The spread of deepfake videos is a growing concern for many people. These videos are created using artificial intelligence and can be very convincing.

They can be used to spread false information or to create embarrassing or harmful videos of people without their consent. Intel has devised a way to detect deepfake videos and help prevent the amplification of these videos in social media as well as being further amplified by the news reporting mainstream media.

This is a welcome development, as deepfake videos can cause a lot of harm. They can be used to spread false information, to create politically embarrassing or personally harmful and defaming videos of people without their consent, or simply to cause confusion. Intel's solution is based on the fact that deepfake videos are often of lower quality than regular videos. This is because they are created using artificial intelligence, which can't yet create videos that are indistinguishable from real ones. By looking at factors such as pixelation, blurriness, and other factors, Intel's solution

FakeCatcher looks instead for authentic clues in real videos, assessing what makes each of us human beings. The detector uses the blood flow signals collected on the whole face and the algorithms translating these signals into spatio-temporal maps. Then, thanks to deep learning, FakeCatcher instantly detects whether a video is real or fake.

Intel develops software it says can detect deepfakes in milliseconds

This AI can recognize if the person is real or not in just a few milliseconds and with 96% accuracy

vocal.media

We seriously need deep flake posts auto-moderated -- like the constant and repetitive $200 a day + pdf attached spam on a first or second post. Seem pretty simple to filter ...