Updated: Apr 28
A number of technologies were introduced for manipulation of images in the 19th century. And these were used in motion pictures later on. Such techniques got improved fast with the digital videos.
Research conducted at academic institutions in the early 1990s helped in developing Deepfake technology. And, the developers fine-tuned this technology in online communities. In recent years, Deepfake has grabbed a lot of attention to be used in hoaxes, financial frauds and fake news.
If we use the technology in a beneficial way, then it may come out to be good for the community. But the government had to limit and watch out for the illegal use of the Deepfake technology.
What is Deepfake Technology?
This technology refers to the creation of manipulated visual content which is generated with the help of sophisticated artificial intelligence. This technology helps in yielding modified images and sounds which appear completely real. In most of the cases, a person in an image or video is replaced with another person.
Deep learning models, which is actually a sub-class in machine learning methods, are used in the creation of Deepfakes. These machine learning methods are generally based on artificial neural networks.
How is Deepfake a big cyberthreat for our community?
Deepfakes can be deceiving, hence they are considered as a threat to us. In 2018, a deepfaked video of Barack Obama was created which had an audio track synced with it and the audio track was by comedian Jorden Peele. Here, people objected to the election or political interference.
It was said that there is not just political or military concern here, rather it had law enforcement, insurance and commerce concerns too. All these concerns were properly validated by an incident that took place in 2019. In this incident, the fraud people used deepfaked audios in order to scam a CEO of a huge company and steal his $243,000. As per the victim’s evidence, the caller used deepfake technology to replicate the German accent as well as melody of his employer and thus fooled him for taking that money.
Despite the examples discussed above, there is still a need for materialization of threats which may get posed by deepfakes. However, this technology is majorly used in creating viral videos or in creating adult content and not in any high tech cyberespionage which causes the security experts, computer scientists or politicians to worry.
Does Deepfake technology pose the similar risk in 2021?
One of the major reasons for Deepfake technology to not be in full-fledged level of scams is the essentiality of knowledge and the method with which they get generated. For creating a real looking or convincing deepfake, it requires the Artificial Intelligence algorithms and deep learning.
Moreover, another factor which limits the use of deepfake is because scammers do not even need such technologies. There are several other low-tech methods which may help them in fooling other people. Then, why will they go for such a high-tech method for scamming?
These reasons do not tell us that this technology will not cause any harm. It is possible that these barriers for creation of convincing frauds diminish with time. As they are getting popular among people, there are several apps also which makes it easier for scammers to create more convincing, cheaper and faster digital deepfakes.
The best way for us to stay away from such cyberattacks or scams is knowledge. Every person who is well informed of the cases will be difficult to be fooled.
How concerned should we be?
As we have discussed, deepfake is not a huge threat to identity fraud. If a person tries to impersonate him as another individual as per the digital identity of that person, in this case digital manipulations cause some sort of threat to the methods of biometric authentication.
But this is not applicable for a fraudster who is amateur. In order to create convincing deepfake images or videos, there must be significant technical expertise along with so much compute power, unlike some approaches like face morphing. So, it is evident that the scammers will need to put in a lot of effort and time in order to develop their capabilities so that they become capable of at least starting with creating convincing deepfake images, audios or videos.
So, unfortunately, we cannot completely rely on this factor of technical barrier so as to prevent the scammers from using deepfake technology for duping other people. As for using any fraud technique or performing cybercrime, people could do this by packaging the code in order to allow others so as to allow others for leveraging
Biometric authentication used to detect deepfakes
However, we do not face such deefakes frequently, we can still detect these deepfakes by using some important techniques.
First of all, biometric analysis powered by AI can considerably help in determining if the video is forged or original. It uses techniques like motion, lip and texture analysis in order to verify the physical presence of the user.
Secondly, we can do it by giving random instructions to the user so that we can know of their authentication. These instructions may include looking in the instructed directions, reading some phrases. These methods will help in predicting their authentication. The users who got to respond wrongly can be taken for further investigation. Also, the manipulation of deepfakes lowers down the video quality significantly.
Lastly, the criminals should be fooled by making them think that we are capturing a deepfake video with the help of a phone, whereas we will use verification systems and fool them. Thus, you can know of the fraud.