Deepfake the most dangerous artificial intelligence software
Created by
Wednesday, August 7, 2019

As many have praised Artificial Intelligence, and extending what technology can reach in finding solutions to many obstacles, we find that those who were warning about it and how dangerous it is to humans are smiling now, because one of the fears they were warning of, even - hidden - has begun Running.



Celebrity deepfakes Victims :

Late last year, in December, some real-life videos of celebrities from Hollywood actors appeared in some abusive videos - p**nographic films - in a first-of-its-kind precedent, for analysts to investigate the credibility of these videos, to find that they were made and developed by Deep Learning, the latest advance in AI.

The Internet explodes with articles about the dangers that this wave can pose, especially in the areas of face change for harassment of innocent people, spreading fake news, and damaging the credibility of video evidence forever.

All agreed that this new technology would be the next wave in the field of false news Fake News, which is full of the Internet, especially social networking sites, to make access to the truth becomes more difficult as we progress in the technical fields.

The first Appearence of the program:

This program was shown through a video fabrication by a Hollywood actress where a short scene opened in a room with a white sofa and a potted plant, and a kind of modern paintings that you can see on some rooms or on the murals of modern art museums. In the room, Gal Gadot was the heroine of "Wonder Women" and "fast and furious", or someone who looked just like her, dressed Eye-catching befitting such videos, showing an unmistakable smile.



That scene, dear reader, is an introduction to a video that appeared on the Reddit online forum, known as DeepFakes, a fake video made with artificial intelligence software, created using a program called FakeApp, and the face of the famous actress was installed on the body of the p**nographic actress. She knows the details of the actress's face better, and she may really think that she is her.

So what are DeepFakes and how does it work ?!

Until recently, the installation of faces or the production of realistic computer videos was an expensive technology, out of the reach of amateurs and designers, and only accessible to large-scale Hollywood producers or researchers with special and distinctive missions.

But with DeepFakes, it looks like it's going to be easy, and an amateur community has started experimenting with FakeApp, a software designed by an anonymous developer that uses an open source software that makes it free and relatively easy to install faces on exciting videos, since a version of the app appeared on Reddit, this software also known as OpenFaceSwap or "OpenFaceSwap GUI" and other version use WinPython Command to run faceswap program written in python programming langquage, Fortunately, this process is not easy and very complex and the program is not so easy to be used from all categories, but will remain a threat in the event of the emergence of sophisticated software easy to use, especially since most of these programs have become almost open source if not fully open.


This version can be considered an advanced version of the famous Photoshop program, but instead of working on still images, the work here is on videos and videos, in the case of Deepfakes the program does not only swap the face but also matches the expression, and if the person's body shape is very similar to the victim's character , A fake video can look very real and cannot be easily distinguishable.

According to the developer of the program, who declined to be identified to avoid legal prosecution, the program is based on multiple open source libraries such as Keras and TensorFlow, which Google provides free for researchers and graduate students, and anyone with an interest in machine learning to collect celebrity faces.

The developer said that he used Google Search to search for images and YouTube videos, knowing that “Deep Learning consists of networks of interconnected nodes that manage accounts independently on input data”, to produce the video in question that has trained the algorithm on videos. Nationality and the face of the actress, and with a few artistic processes a professional fake video was produced that could not be easily distinguished.

Unlike applications that use face swapping for funny purposes, FakeApp uses artificial intelligence to complete three key steps: Alignment, Training, and Merging.

Instead of placing one face on another in real time, FakeApp uses hundreds of still-frame images that are pulled from a video or search engine, digging through all those images, identifying faces, analyzing how it is illuminated and what expressions it makes, etc., Once the program understands the faces in which it works, it can use all the "knowledge it has acquired" to link one to the other to produce a video or visual clip of the person concerned.

What are the serious aspects of this new approach ?!

With the big explosion of information, not everyone recovering from the Fake News revolution that has swept the Internet, and technology companies are trying hard to fight it, we find that this new technological development will add more burdens to the simple user in distinguishing the real from the false in the surfing the Internet.

Faking videos and videos is no longer that difficult, regardless of the technical challenge.Any amateur programmer or someone who wants to harass you or defame you online will need enough pictures of your face online to create a clip, install your face in some videos, and post it online.

Currently, there are a number of apps on the Google Play Store, and Apple App stores that allow you to swap your face for someone else. Even Snapchat has one face swap app with some filters.Many people use these filters to create funny videos online, but none of This looks real like the one created by Deepfakes, and here it is dangerous.

Not only are some of the videos defamatory or offensive, it can be frightening to penetrate the privacy of users with their images spread on the Internet, and on social networks.

Is there a law criminalizing this technique ?!

“There are not many legal ways to redress the victims of this new technology,” says Jonathan Masur, a professor of patent law and technology at the University of Chicago Law School, even for ordinary citizens.

"There is a copyright claim if you take 'shots' yourself, there is a defamation claim if someone tries to say it's actually you, and if you're a celebrity, there's a right to claim publicity if someone tries to make money from this video."

“But each of these is just a narrow slice of what is happening here. The case won't cover the vast majority of these cases, you can try to show a case of some form of defamation if you're attacking someone's reputation, but it's also very difficult, because you are Definition Don't pretend you are posting an offensive picture of that person. ”

As with many emerging technologies, the most terrifying part may be invisible even in this technology, and we all remember when Facebook was launched from the university campus, few people expected it to turn into what it is as the first source of false and fabricated news broadcasts.

Like the wave of “fake news” that has intimidated the Internet and users, the depth of such technologies is another aspect of the Internet's ability to breach our real reality, and if every video is likely to be fake, why do we think anything else is real ?!

This article has
5 points
Give points
1 2 3 4 5

I like

You like

Report

You have reported
Reasons
Cancel
0 comments






To Advertise Here Contact us