They are not (although they are related) the well-known “fake news”, although its meaning translated from Spanish is “ultrafalso”, An acronym for English formed by the words fake (falsification), and deep learning (deep learning).
It is a technique of artificial intelligence that allows you to edit fake videos of apparently real people, using unsupervised learning algorithms, known in Spanish as RGAs (antagonistic generative network), and existing videos or images.
Its use in politics is growing to present candidates or officials making inappropriate statements, and it will undoubtedly be a great barrier to overcome for the truth in the near future. The end result of this technique is a very realistic, if fictional video.
How could it be otherwise, the first convincing trials of “deepfakes” were with pornographic videos allegedly starring US movie stars. The first “deepfake” to the scene in the movie “Forrest Gump” is considered when actor Tom Hanks appears (in character) alongside President Nixon for a few seconds of an original filming presented as a newscast.
THE PRINCESS WHO DID NOT DIE
Another famous example of this technology brought to the cinema was seen in the last film in the “Star Wars” saga, in which actress Carrie Fisher (Princess Leia), who died during filming, continued to “act” in scenes that had yet to be recorded.
Norwegian actress Ingvild Deila played Princess Leia in “Rogue One,” a Star Wars story, although her face did not appear on the big screen. Thanks to the same techniques that are used in the “deepfake”, it was possible to see a very young Carrie Fisher, although she was about to turn 60 years old.
Shortly thereafter, Fisher passed away during filming and the same techniques were applied again. Even so, the world of cinema debates the legitimate position and opportunities that this technology offers to the production of audiovisual content and how much it affects the generation of job sources and caches.
Thanks to RGA algorithms, photographs can be generated that appear authentic to a non-specialist observer. It is clear to understand why the use of this technology has been directed primarily at public figures. Due to the vast amount of audiovisual content available with the image of these people, it is easy to feed a database of the model to be able to create compelling mock images.
A MESSAGE FROM BEYOND
Like all technology, it is neither good nor bad in itself, it depends on the use made of it. The imagination last week overcame the mental frontier on how to use a “deepfake”. A young man was brought back to life with the consent of his parents after being killed in a school shooting. Joaquin Oliver, 17, was killed in a shooting in Parkland, USA, in 2018.
Now it appears on the networks and TV with a powerful message: “The elections this November would have been the first in which I could have voted,” the digitally revived spectrum tells the camera, with gestures that imitate the young man’s body language, “But I will never be able to choose the world in which I want to live”, sentence.
Joaquin’s parents have spent the last two years completely committed to promoting legislative changes that curb gun ownership, one of the main contemporary evils of the American nation. Theirs is the path taken by many parents who lost children in recent massacres, but with “The unfinished votes” they have gone a step further, putting words that they never said into their child’s mouth and body, trying to shock and move the eldest. number of possible consciousnesses.
In an interview, they assure that the process of making the video was “the most difficult experience we have had since we began to fight against gun violence after the murder of our son”, but it is something that will help change the role of Joaquin from victim to activist, they allege.