Is deepfake dangerous?

I’ve been reading about deepfake.

I think everybody can see the potential danger if such technology is used by people with dark intentions, such as hackers and scammers.

However, it might be a bit irrelevant. The technology is here to stay, so, instead, we should probably learn to better detect deepfake.

Besides, while there are algorithms to generate deepfake, other algorithms can detect them. Roughly speeking, deepfake algorithms often leave artifacts behind, even if you can’t spot them with the naked eye.

Source: Unmasking DeepFakes with simple Features.

I’ve been playing with libraries, and I found this one, which helped me A LOT for this demo.

I was surprised by the easiness, though. You can generate pretty realistic human faces for free in seconds.

See Also