- Corresponding Author: Christos Beretas. PhD Candidate in Cyber Security at Innovative Knowledge Institute, Paris, France.
- E-mail: cberetas@ikinstitute.org
- Policy Paper Title: DeepFake: Another one Cyber Threat
- Keywords: deepfake, deep-fake, deep fake video, deep fake photo, algorithms
Abstract
In recent years, with the innovation of information technology, they have introduced various kinds of new applications, many of which were created for satirical purposes, but have quickly moved to the opposite side, and skeptical people are using DeepFake technology to gain profits and publicity, breach of personal data and fulfilment of strategic goals whether they come from political opponents or the interests of third parties, DeepFake technology quickly became abusive and exploitative. It has led people who virtually involved in photos and videos to lose their jobs or even lead to social exclusion. DeepFace is a profitable business around the world, as long as reflect on the extent of the problem, as soon as 2019 the fake videos have doubled and all research is showing an anodic trend. Technology is a flagship, the use of technology against humans makes technology the biggest enemy.
Introduction
DeepFake tends to become the main threat worldwide because it is not easy to detect a fake image or a fake video. The main reason for DeepFake is to humiliate famous people, which is why most DeepFakes are pornographic videos aimed to humiliate famous people. Of course, DeepFakes is used in other areas not only in pornography but also in government agencies around the world as a lever to oppress the morals of third parties who are in opposition or who are leaders of third countries. DeepFakes can also elect a government, a well-established DeepFakes-based communications strategy could mislead and stimulate a significant share of voters, as people tend to believe what they see, this is something that is difficult to reverse. Many times fake videos have been presented in courtesies by distorting the facts, neither have the cynical activities of corporate executives or even blackmailing corporate executives in order to misappropriate money from these companies.
Artificial Intelligence
The evolution of artificial intelligence has undoubtedly heightened DeepFake videos, nowadays it is easier than ever to create DeepFake videos. Technologically advanced countries with the most advanced artificial intelligence capabilities involving access to large volumes of data will create enormous problems in the technological world as the creation of deepfake can pose a global threat not only to the digital world but also to the real world, due to false pretexts, triggering outbreaks of violence, security problems, social and political problems.
Nowadays, thanks to artificial intelligence, these kinds of applications have been significantly improved, so that users are able to pick a video on their computer from whichever medium then, they download from a free library from a neural network and associate the person they want, then the ability to replace the face of the video with the simple push of a button.
The following Figure 1.1 illustrates the process of creating DeepFake.
 |
| 1.1 – DeepFake Production Steps |
As shown in the above diagram, the Encoder extracts facial image features and the Decoder is used to redesign the facial image. The key to changing the two pairs of photos is the existence of an Encoder and Decoder, because each pair is used to parse a picture and to parse a picture both parameters are used in common. The above technique allows the Encoder to learn the facial features and identify the similarities that exist where the new facial reconstruction is then performed.
As artificial intelligence evolves, the potential for more profound DeepFake video increases, as the proficiency of DeepFake video increases, so do people's trust less social media networks and the media generally as they realize how fake manipulative news is everywhere, DeepFake as the number one digital threat. The reason is simple, uncontrollable use of artificial intelligence that in quotation marks could be called (immunity) enables people in power to cause discomfort, financial impact, defection, hatred, create political tensions, manipulate citizens to the extent that they can even lead them to civil or non-civil wars.
It is a global threat as anyone can manipulate a third party, thanks to the freedom of using artificial intelligence such applications become more accessible and easy to use, allies in the process are social media that reproduce without examining fake content and affects the honour and reputation of the citizens
Literature Review
DeepFake is a global threat for all areas, including privacy, personal life disorder, hostilities, political unrest and much more. Research done in recent years have shown that there are 14,678 videos online or 96% of them where these DeepFake videos are pornographic, which means either that they have a purely intentional purpose to create this fake video to discredit this person or to satisfy their personal ambitions and fantasies. The remaining 4% includes videos created for political manipulation and propaganda, as well as satirical videos. 90% of DeepFake videos feature popular actors, artists and politicians. In most cases of pornographic content are actresses who are embodied in pornstar bodies.
Conversely, men embodied in images and videos do not come from the porn sector but mainly from the political and artistic sector. 2/3 of pornographic videos are made up of people who live in western countries, while 1/3 show people who do not live in western countries. In conclusion, in most cases DeepFake video creation is pornographic content with a Western faces as a protagonists. With the advent of Generative Adversarial Networks (GAN) which are nothing more than two neural networks that include a synthesizer, a generator, and a detector where they then create images and videos with the advantage of repeating the whole process until they are attained. of the end result. It is a very good and effective method of creating DeepFake.
Research on DeepFake video refinement are ongoing, the video production industry is profitable, and of course anyone can order the video of interest for the purposes of interest.
Figure 1.2 below presents the 3 most popular DeepFake creation techniques.
 |
| 1.2 – DeepFake Generation Techniques |
On the other hand, DeepFake creation tools have made their appearance, which sleeps from low to high cost, easy and versatile to use. As time goes on, more and more DeepFake creative tools will make their appearance due to the impressive result the attraction is great even for political purposes.
Figure 1.3 show the increasing number of websites that host fake video and it is precisely the increasing number of fake video.
 |
| 1.3 – Increase of DeepFake |
 |
| 1.4 – GAN Papers Increase |
Survey of how threatening DeepFake is in 2019 revealed that 88% of Americans believe DeepFake causes harm, as opposed to 47% of Americans who believe there is a slight chance of falling victim of DeepFake. The reasoning behind the people who participated and answered negative (47%) responded with the reasoning that DeepFake is affecting public persons, the purpose is to pervert a public person and not a non-public person.
No one would be interested in watching a DeepFake video showing unpopular people. 1 in 10 men in the above research believes DeepFake should be legal but with restrictions. 25% of Americans believe they are afraid of falling victim to DeepFake, which includes more social and popular people who are heavily involved in society. 21.6% of Americans do not trust the media believing that they are manipulated. Over 60% of Americans believe that DeepFake can be used in elections to influence voters. 82.1% believe that in the next decade they will be much more skeptical of the news they are reading online.
It is increasingly clear that DeepFake is a global digital threat, lacking a legal framework with a strong growth trend and uncontrolled action. The following Figure 1.5 (blogs.forbes.com) shows a well-known search engine visitor searches of the words “deepfake”, “deep fake”, “deep fakes”, “deep fakes”.
 |
| 1.5 – DeepFake Search Percentage |
Key Consideration
As discussed in the previous chapter, there are several techniques for creating fake images and videos, the research presented undoubtedly showing the seriousness of the problem and how uncontrollable it is, is a global digital threat to privacy, to humans, and ethics. So far, it is worth noting the following assertions that are all essential.
• There is a lack of legal framework worldwide, and this may well suit some people around the world who want to take advantage of technology for personal prosperity.
• Research about DeepFake's are fast-paced, and research teams are trying to continually improve the quality of the output.
•There are confusion and abuse of people who are alleged to be involved in videos without their will, many times that these videos are offensive and violate human dignity.
• Politicians lack the legal framework to use technology to kill opponents.
• Governments use technology to expose third-country governments to blackmail.
• Media manipulation, proprietary material designed to influence citizens, members of corporations and entrepreneurs.
• They cause social exclusion to people who are involved in either videos or photos that have been targeted and believe that they have been tricked into doing so as a result of their financial destroyed.
• Offending morals and desires, people are reminded of being involved in videos that offend their faith, morals and habits.
• When DeepFake is used for social killings and hatred to launch hostilities against individual people or communities.
• When a global digital threat arises without the corresponding interest in countermeasures.
Challenges
The rapid evolution of artificial intelligence, coupled with the increased zeal of research teams to refine the result, the global community has to address the following challenges.
• Detection and isolation of fake videos.
• Creating a legal framework.
• Restriction on the use of artificial intelligence with terms.
• Protection of personal life.
• Creating a legal framework for fake videos in the media.
• Working groups to create countermeasures.
Recommendations
DeepFake undoubtedly has an impact on society, democracy, ethics and causes a worldwide digital threat. DeepFake's detection methods are not sufficient and unsatisfactory. A number of suggestions are presented below.
• Initiate research into the use of Blockchain to detect fake videos and photos.
• Creating a legal framework that will be able to determine what is permitted and not and to what extent the affected can legally turn.
• Create research teams that will analyze DeepFake's creation algorithms and try to refocus their creation techniques.
• Creating algorithms that could parse the structure of a fake video to prove whether the video was original.
• Creating a database of real videos so that the comparison with the real video and a fake one could be compared.
• Create an algorithm for analyzing video and photos at the code level to detect suspicious points.
• Create an algorithm that focuses on the images presented to examine their distortion.
• Create an algorithm to detect fake photos that will control the image density for lesions.
• Search for a photo on the internet looking for a duplicate of another name on the internet.
• Creating algorithms for Video in frame by frame level analysis to detect cut points.
• Creating a face detection algorithm that detects changes in facial features during video playback.
Conclusions
People's confidence is slowly starting to diminish in the media. As they are beginning to believe that what they see or read may not be what they actually happened. DeepFake can cause discomfort and hatred for people or to the people who are targeting in one community, increasing the chances of violence, conflict and racial discrimination. DeepFake is a global digital threat and has come to stay, researches for fake video refinement are on the rise, and the DeepFake technology of either video or images is becoming increasingly accessible.
DeepFake is a global digital threat because the creators of such videos and photos do not need to disseminate this material on social media, behind this design there are process flow charts choosing media they know in advance to disseminate fake news to the fullest extent. Therefore, there is a network of people who are not liked is karatom, such a tactic is shown by intelligence services acting on behalf of government officials acting against third party leaders, the nation itself or even sabotage.
Quite often videos are presented in courts and based on these videos are targeted against people which may be unfair, there is no credible and valid mechanism to verify that the videos presented in courts are real, and yet these videos are judged people and lives and families are being destroyed. With all that has been clearly stated in this policy paper, DeepFake is a global digital threat based on lies and fraud, mercy-free fraud without restrictions.
As has been mentioned above, it is very difficult to detect fake images and videos, due to the continuous refinement existing in artificial intelligence applications. Algorithms have been designed to detect fake images and videos, which in practice have been found to be inadequate. The reason for creating this policy paper is to report the real global problem that is equivalent to a global uncontrolled digital threat, and to suggest to policymakers some algorithms and techniques for making fake videos and images easier to find.
References
- Figure 1.1 miro.medium.com
- Figure 1.2 lilianweng.github.io
- Figure 1.3 technology-info.net
- Figure 1.4 technology-info.net
- Figure 1.5 blogs.forbes.com
- Literature Review zdnet.com
- Research / Analysis The State of Deepfakes: Landscape, Threats, and Impact (regmedia.co.uk)
Do, not Miss Anything!!! Subscribe to our Email Updates to get the latest Articles in your Email instantly.
Your articles are very well written and unique. king99
ReplyDelete