Deepfake - an Amusement as Well as A Huge Danger on The Internet

Fake videos of certain public figures who voluntarily reveal something that causes outrage are increasingly appearing on the Internet. Most of this is blackmail, misinformation, and so on.

In the article, I would like to explain to you what a deepfake is and how you can generally expose such scammers

What Exactly Is a Deepfake?

The word deepfake is made up of the words deep learning and fake and is basically an artificial intelligence that manipulates videos and audio tracks in such a way that individual faces are exchanged or edited in such a way that they look real to the viewer.

To make the whole thing clear, I would like to explain it using an example: The YouTube user “Ctrl Shift Face” has edited the face of Macaulay Culkin, the actor from “Kevin - Alone at Home” so that Sylvester Stallone can be seen in it . This may seem funny at first, but deepfakes are also used on more serious topics.

The founder of Facebook, Mark Zuckerberg, says the following in a deepfake: “I have control of all your data. Whoever masters the data also masters the future ”. This caused unrest among many viewers, as the video was manipulated so well that it looked real. It turns out that the artists Bill Posters and Daniel Howes have collaborated with the advertising agency Canny Al. They wanted to demonstrate to society how advanced the technology is, because there is currently a competition between scientists and fraudsters.

Deepfakes can now be produced with simple programs within two days. You need a good graphics card for this. In fact, 95% of deepfakes are exploited for pornographic content. Indian journalist Rana Ayyoub was blackmailed with fake porn videos. Governments in California have already passed two laws that do not allow human faces to be used in pornographic material or on political topics.

How to Protect Yourself from Deepfakes?

One of the most important things you can do is educate people: Make sure you tell your fellow human beings how dangerous deepfake is, because once they are educated, they will usually be less likely to fall for such scams.

You should also analyze people or faces in videos more precisely. Pay particular attention to the lighting of the face, with a deepfake it is often the case that the background lighting does not match the face lighting. You should also pay attention to facial expressions. Pay attention to strange blinking of eyes and lip synchronicity.

How Can We Protect Ourselves from Deepfakes?

Governments are slowly starting to address the threat posed by deepfake video. For example, two laws were passed in California last year that made deepfakes illegal. The first, AB-602, prohibits the synthesis of human faces to create pornographic material. The second, AB-730, prohibits tampering with images of political candidates within 60 days of an election.

But are these measures enough? Fortunately, cybersecurity providers are constantly developing new and better detection algorithms. These algorithms analyze the video image and find the smallest distortions caused by the forgery process. Current deepfake synthesizers, for example, model a 2D face, which is then distorted to create the 3D perspective in the video. The direction in which the nose is pointing is a good indication of this.

Deepfake videos are still at the stage where you can spot the signs for yourself. Look for the following characteristics of deepfake videos :

·         Jerky movements
·         Changes in lighting from one frame to the next
·         Changes in skin color
·         Strange or no blinking at all
·         Lack of synchronicity of lips and language
·         Digital artifacts in the picture

But the better deepfakes get, the less you can rely on your own eyes and should seek the help of a good cybersecurity program instead.

Current State of Anti-Fake Technology

For example, some new technologies help video authors authenticate their videos. A cryptographic algorithm can be used to insert hashes into the video at fixed intervals so that it can be recognized immediately if the video is falsified. AI and blockchain can register forgery-proof digital fingerprints for videos. This works in a similar way to watermarks in documents. The difficulty with video, however, is that the hashes must persist when the video is compressed for different codecs.

Another method to prevent deepfake attempts are programs that insert specially constructed digital artifacts into videos. In doing so, they cover up the pixel patterns used by facial recognition software. This slows down deepfake algorithms and produces poorer results, which reduces their likelihood of success.

Good Security Practices Are the Best Protection

But technology isn't the only way to protect yourself from deepfake video. Good basic security practices can also be effective in preventing deepfakes.

For example, many deepfake and other fraudulent actions can be prevented by integrating automatic checks into every process for paying out funds. More options:

·         Make sure employees and their families understand the deepfake problem and its dangers.
·         Train yourself and others to recognize deepfakes.
·         Build good media literacy and use quality news sources.
·         Use basic protocols (“trust but verify”). A little skepticism about voice messages and videos won't protect you 100 percent from fraud, but it will help you avoid many pitfalls.

Remember, when hackers start using deepfakes to break into home and corporate networks, basic cybersecurity best practices are critical to mitigating the risks:

·       Regular backups protect your data from ransomware and offer you the opportunity to restore damaged data.
·       By using different and secure passwords for your various accounts, you prevent a successful attack on one network or service from affecting everyone else. If someone breaks into your Facebook account, they shouldn't be able to get into your other accounts afterward.

·       Use a good security package, such as Total Security Software to protect your home network, laptop, or smartphone from cyber threats. One such package includes antivirus software, a VPN to prevent WiFi connections from being hacked, and protection for your webcams.

What Does the Future of Deepfake Look Like?

Deepfakes are always evolving. Two years ago it was still very easy to recognize deepfake videos: by the bumpy movements or by the creepy eyes that never had to blink. But the latest generation of fake videos has evolved and adapted.

There are now an estimated 15,000 deepfake videos in circulation. Some are just for fun, others are meant to influence your opinion. However, since it only takes a day or two to create a new deepfake these days, that number is likely to grow dramatically.