Samsung develops technology to create fake videos from a single photo

Researchers at the Samsung Artificial Intelligence Center (AI) in Moscow have created an algorithm that can generate videos using a single image.

This development has raised some concerns among technology experts and commentators, who see it as a worrying step in facilitating the creation of false content.

In an article published in the ArXiv pre-print journal and in an accompanying video demo, the algorithm creates a video using a single still image, such as as the painting of Mona Lisa or a picture of Salvador. Dali.

via GIPHY

The video can be created with the help of a single image, but the more the number of images used is high, the better the quality.

A sample of 32 images produces a precision video close to reality.

Current artificial intelligence systems typically require an algorithm to scan large sets of body and face data before they can produce an animated image based on it.

With this new technology, the creation of fake videos

The Samsung algorithm was formed on the public database VoxCeleb, which contains more than 7,000 celebrity images from YouTube videos.

This algorithm recognizes the common characteristics of the face and body of a person, as opposed to

This method also means that the technology is applicable to non-celebrities and can be used by anyone, even by those who died a long time ago

Artificial intelligence is currently able to produce only "talking head" style videos from the shoulders.

The skeptics of deepfake technology, as it is called, are afraid to use it to spread false information and false news or to steal people's identities.