New world with deep fake
The Internet is now available in almost every iron, it becomes more and more difficult to find a point on the map, where 4G network would not be available. Broadband Internet is primarily HD quality video on common platforms, which are gradually replacing news feeds for us, analytics and just entertaining reading. Besides, it is an instrument of influence on billions of people, allowing to create the necessary public opinion at the moment. Deep fake technology may be at the peak of these games with the population, which has already proven its ability to turn celebrities, such, like Gal Gadot, Scarlett Johansson, Taylor Swift, Emma Watson and Katy Perry, the stars of the porn industry. Deepfake is an algorithm, allows you to simulate the behavior and appearance of a person on video. The technology got its name from a combination of deep learning ("deep learning") и fake ("fake"). Deep fake is based on the notorious neural networks, operating on a generative-adversarial principle (Generative Adversarial Network, HOWEVER). Algorithms, programs included in the database, constantly compete with each other in two processes: training on the submitted photographs in order to create a real substitution of the face for copies and exclude unsuitable options until then, until the machine itself starts to confuse the original and the copy. This complex scheme is the main goal of deep fake - the creation of false photos and video content., in which the face of the original is replaced by another. For example, charismatic US President Donald Trump may well take the place of any odious leader of the 20th century and carry outright heresy to the masses from the podium. Now, in one of the generated deep fake videos, former President Barack Obama allowed himself to use obscene language against Donald Trump.
Of course, at first, deep fake ads were exclusively viral - the faces of Hollywood actresses were implanted in plain plots of porn videos with all the ensuing consequences. Or, eg, actor Nicolas Cage unexpectedly becomes the main actor in episodes of the most iconic films of our time. Some of these creations are shown on video and, honestly, many of them look a bit clumsy.
And here's the hip-hop artist Cardi B, who appeared on the Jimmy Fallon Tonight Show as actor Will Smith, looks quite convincing.
And here is the original.
Crafts look good on the Ctrl Shift Face channel. For example, Sylvester Stallone tries on the main role in the "Terminator".
IT analysts are already saying, that "fake" technology could become the most dangerous digital technology in decades. In addition to this technology (and based on it) experts from Princeton universities, Max Planck and Stanford built the Neural Rendering app. He has an even more dangerous task - to “force” the image of a person with his voice to pronounce any text in any language. It only requires a 40 minute video with audio, during which the neural network will learn to work with the voice and facial expressions of a person, transforming it for new verbal expressions. Original idea, of course, positioned as good: film studios will become the main consumer of Neural Rendering, those who want to reduce the time for shooting acting doubles. However, it soon became clear to everyone, that in the near future, virtually any user can generate video fakes on their laptop, that will make your hair stand on end. The second branch of the "fake" program was the application DeepNude, capable of quite realistically "stripping" naked in the photograph of any woman. In the early days of the service, the volume of dropped requests was so high, that the developers, in fear of lawsuits, announced its closure. But hackers hacked the resource, and now everyone can use DeepNude. Of course, service is trying to limit access, but everyone understands, that this is just a temporary measure.
Deep fake turned out to be a good tool in the hands of scammers. One British energy company was robbed on 220 thousand. euro, when a “clone” of a manager from Germany got in touch with the financial department. He asked in a simulated voice to urgently transfer money to an account in Hungary, and business partners did not have a single reason not to trust the subscriber. Of course, now it is problematic to massively and qualitatively fake video sequences - state regulators constantly block resources with fakeapp and facefake, and the power of computers does not yet allow you to quickly synthesize video. We have to delegate this work to remote paid servers, who require thousands of photographs of the original personality and the "victim" to work.
The beginning of war
Deep fake technology may leave actors out of work in the future - the entire industry may well switch to “cartoon” movie characters, many of whom technology will raise from the dead. But it's more of a dream of the future, as numerous unions and a basic lack of computing power pushes the prospect forward several years. Although already now in the film "Rogue One: Star Wars director for one episode "resurrected" actor Peter Cushing, deceased in 1994 year. rumor, that the famous James Dean may appear in the new film about the Vietnam War. Neural networks help actors, who are already in years, look in the frame 10-20 years younger, for example, Arnold Schwarzenegger and Johnny Depp. On average, the world now generates at least 15-20 thousand. deepfake videos every month, most of which appear on the Internet. Russian programmers are trying to follow the global trends - Vera Voice in July 2020 year will offer fans of Vladimir Vysotsky's work to chat with a digital copy of the singer at the Taganka Theater.
Everything goes to that, what video- and photographic evidence will cease to be effective arguments in court proceedings, and the global video surveillance system will be a waste of money. No one will trust CCTV footage. Where is the guarantee, that this is not a synthesized pacifier? In political propaganda, deep fake is already becoming the most powerful lever of influence on the opinion of voters.. California in October 2019 year became the first state, who generally forbade posting videos with political candidates online for 60 days or less before the election. For violating this law 730 facing criminal liability. Now several more states have joined this initiative., and January 2020 year in China it will be forbidden to publish synthesized deep fake photos and videos without a special mark. by the way, now one of the most effective methods for determining a fake by eye is the lack of natural blinking in synthesized characters.
Stallone as Terminator
Now imagine, how much the development of deep fake technology (but development cannot be stopped, as long as there is demand) will change the concept of truth and lies. Especially when government agencies adopt the novelty. Any synthesized video about another espionage exposure, if used skillfully, can become at least a pretext for imposing another set of sanctions or closing a diplomatic mission. And there will be no need to stage chemical attacks on civilians, to authorize a massive missile attack on a sovereign state. Acts of genocide, consequences of the use of weapons of mass destruction, defiant and insulting behavior of the first persons of the state - all this deepfake bouquet in the right situation can justify the start of another military adventure in the eyes of voters. And when the guns speak and the rockets fly, no one will really remember, what revealing video started the war.
Right now there is no clear answer., what to do about this disaster. Best Algorithms, sharpened to expose deep fake, can only guarantee a 97% chance. Currently, every progressive programmer can take part in the Deepfake Detection Challenge, which in September 2019 announced by Facebook. Prize fund in 10 million. dollars will go to, who can develop the algorithm, allowing with a 100% guarantee to recognize fake videos. How fast will the response from underground deep fake developers follow?, only guess.
Yevgeny Fyodorov
