In order to counteract the deepfakes, fake videos have been “artfully” created, with the help of which fake news can be distributed even more efficiently. Google is also active in this area. In fact, the technology giant has just released a so-called “large set of deepfake visual data,” a database of videos created with deepfake technology that can be used as a comparison set to identify other videos.
This further step makes it clear that even the technology giants and not just governments, politicians and all those who are publicly known are worried about this new technique, which makes it possible to create almost real videos with all faces and people.
Various analysts expect that by the end of next year the videos will reach such a high level of quality that the human eye will no longer be able to distinguish them from an actual video.
This is precisely why special software needs to be implemented that is likely to be more efficiently integrated with machine learning techniques to identify these fake videos in an automated way. According to Nick Dufour of Google Research and Andrew Gully of Jigsaw (one of the many companies in the alphabet universe), the Google database was created in collaboration with willing payers.
Hundreds of videos were recorded using deepfake technology, publicly available methods and software, and real videos that can be compared with the first ones to achieve a better identification technique.