It’s like the talking portraits in Harry Potter.
It’s now easier than ever to create fake videos thanks to Samsung’s latest innovation: artificial intelligence software that can generate moving scenes from a single photo.
Before this new discovery, previous software could only create a slightly obvious fake video after gathering a large number of photos of a single person together. The software then determines similarities between the facial structure in the photos in order to map that certain face onto a video.
With Samsung’s new AI, however, there’s no more need to get multiple photos. In fact, you can even use just one photo and the software can already utilize that to ‘paste’ the face onto the person on the video.
So far, the quality of this new AI’s work is yet to be confirmed, as Samsung is still testing the software out. The quality also depends on several other factors, such as the lightning of both photos and whether they’re different or somewhat similar. Obviously, if it’s the latter, the software will have a much easier time.
Authors of the study on this AI, however, are aware of the potential negative consequences that this new software may cause. One of the biggest issues might be its possible use in the online pornography industry.
According to a statement on YouTube, they say, “We believe that telepresence technologies in AR, VR and other media are to transform the world in the not-so-distant future.”
“We realize that our technology can have a negative use for the so-called ‘deepfake’ videos. However, it is important to realize that Hollywood has been making fake videos (aka ‘special effects’) for a century, and deep networks with similar capabilities have been available for the past several years.”