Deepfake videos whose algorithms had not been previously studied were used in DFDC final

In the previous article, we mentioned the results of the first global deepfake detection model development competition, which was hosted by Facebook and Microsoft and lasted for more than six months. In the final stage of the Deepfake Detection Challenge (DFDC), in which 2114 teams from each continent participated with more than 35,000 deepfake detection algorithms, the model that won the competition performed the deepfake detection with a 65.18% accuracy rate. This result revealed that the developed AI-based cybersecurity algorithms were able to catch up to two-thirds of all kinds of deepfakes.

However, the average performance of the algorithm that completed the first public stage of the competition in first place achieved 82.56%. This is because, in the first stage of the competition, the competitors were allowed to access the data set, which consisted of more than 100,000 videos, and to train the models they developed on this data set. In the final stage, competitor deepfake detection models were tested on a different data set of 10,000 videos, defined as “black boxes,” that had not been shared with competitors. Moreover, these videos had been modified to make detection even more difficult by using different deep frame rendering models, image enhancement techniques, and additional magnifications and distracters such as blur, frame rate change, and over coats. For this, they enlisted the support of experts from the DFDC academic stakeholders, Cornell Tech, MIT, Munich Technical University, UC Berkeley, Albany, SUNY University, Maryland University, Naples Federico II, and O xford University. This is how they measured the abilities of competitor algorithms to overcome the difficulty of generalizing from known samples to unknown samples. It has become very clear that the developed deepfake detection algorithms have not been as successful in detecting deepfake samples that they have not been trained for, as well as the data sets they have been trained in.

Does not the R&D of deepfake detection models include reverse engineering?

You have already seen that in the process of developing AI-based deepfake detection models, an important part of R&D studies is the effort of cybersecurity providers to solve the algorithms of deepfake tools. One reason for this may be considered to be the innovation initiatives that will lead to the beneficial use of deepfake technologies in areas such as education, health, sports, and communication. However, on the other hand, developing the determination methods in accordance with their working principles by solving the structure of the existing deepfake algorithms is one of the main objectives of R&D. In this respect, R&D for deepfake detection algorithms also serves as a reverse engineering activity for deepfake technologies.

The average performance of 65.18% achieved by the DFDC winner should not be the targeted hit rate for the most successful deepfake detection tool. However, Facebook, which hosted the competition with million-dollar awards, did not even expect this before the competition. Mike Schroepfer, who has been serving as Facebook’s Chief Technology Officer (CTO) since 2013 said they were pleased with the results achieved. Schroepfer stated that the results will guide their future work and said, “To be honest, the competition was more successful than I expected.”

Another interesting determination of Facebook’s CTO is that deepfakes are not a major problem for Facebook at the moment. Mike Schroepfer told journalists that they must be prepared to determine such content in the future. Facebook announced in January that it banned deepfake content, although its scope is controversial. Schroepfer emphasized the importance of being prepared and said, “The lesson, learned for the last few years the hard way, is that I have to be prepared for many bad things that will never happen.” It is likely that nobody except Facebook management sees the deepfake threat as a possible future problem anymore. The fact that Mark Zuckerberg, who is known to be against content restrictions on Facebook, has been among the victims of deepfake proves that it should not be so optimistic.

Facebook also announced that it will contribute to developers who do not participate in DFDC to develop an effective deepfake detection model to protect the online world and, of course, digital platform investments from future danger. Accordingly, it does not find enough to publish the special data set, which forces competitor algorithms only in the final stage, with open-source code after the competition. The CTO of Facebook announced that successful deepfake detection algorithms ranked in DFDC will be accessible to all developers working in this field with open-source code.

The code of the models, ranked in DFDC, is open, but Facebook’s codes are no

Facebook seems to be investing in the future of deepfake detection, along with other online world giants. However, it does not place its hope only in the models developed by the DFDC participants. In his assessment for journalists on the DFDC results, Schroepfer emphasized that Facebook has currently developed its own deepfake detection technology apart from this competition, and says that “We have deepfake detection technology in production, and we will improve it in this context.” Moreover, Facebook also announced that it included this deepfake detection model, which directly contributed to its development in the race at DFDC, excluding awarding.

The CTO of Facebook dropped the real bomb in relation to the deepfake detection model they have developed and explained that Facebook will keep its detection technology confidential to prevent reverse engineering. However, Facebook does not show the same sensitivity for the most successful deepfake detection algorithms that ranked in DFDC. Schroepfer emphasized that they will keep their algorithms confidential and admits in a sense that they will provide reverse engineering opportunities for the algorithms that were developed in DFDC by publishing them with open-source code.

Today, the race between the deepfake technologies to not be caught and the deepfake detection models to catch up is a race that is not conducted under equal conditions. As deepfake detection algorithms improve, reverse engineering will become more crucial for malicious deepfake manufacturers. It may seem that to be an open-source, the global corpus that comes to the forefront in the DFDC organization in this field is intended to contribute to the development of deepfake detection algorithms, and it serves for this. However, this strategy will also help deepfake tools to overcome detection algorithms more easily. Does not Facebook explicitly explain that it is for that reason it avoids to publish the algorithms they have developed?

Due to the COVID-19 pandemic, the digital lifestyle is being pumped into society even more. Artificial intelligence (AI) technologies, such as deepfakes, which are known to pose various dangers for the online world, are shared on the internet for everyone to access and use easily. Then fear is released to the online world. Warnings are issued against disinformation created by fake synthetic sounds and images that are convincing enough to make facts suspicious and blind intelligence. Additionally, global contests are organized to develop deepfake detection models with great expectations and millions of dollars, led by the leaders of the online world such as Facebook and Microsoft. The result obtained after months of waiting is not the success of catching the deepfakes of the developed models. To achieve this success in the future, the existing global R&D will be made available for common use. Fortunately, there are no deepfake production tools that can only be used with malicious intent.

Deepfakes are constantly mutating

Deepfake technologies will undoubtedly have many useful applications in areas such as health, sports, communications, and marketing, as well as other innovation opportunities from AI. This legitimizes the developed deepfake technologies, and so, nobody is in a position to demand that deepfake technologies, which are expected to cause a great threat that cannot be compared with their possible benefits, to be blocked by prohibitive policies. This would be an act against freedom of expression. However, experts agree that it will take time to develop effective detection tools against malicious, dangerous, and harmful deepfake production, and perhaps the war of these algorithms will never end. This is because new technologies, which make deepfakes more believable every day, are constantly being developed in this field. In other words, deepfakes are constantly undergoing technological mutation. Deepfake detection models, on the other hand, are expected to be a miracle drug that quickly ada pts to every new deepfake technology and catches all types of deepfake and prevents their harmful effects.

Mosquito sprays will only work if you remedy swamp

For centuries, in the perfect ecosystem of nature, whose secrets science has been trying to unravel, every being forms an extraordinary part of the sensitive gear. In the life cycle, every creature needs another because it is the only way that natural life can continue. Even blood-bearing mosquitoes, the carriers of deadly infectious diseases, are the food source of many different creatures, from fish species to various birds, bats and spiders. Science advocates that mosquitoes are not a key species for ecological balance. However, instead of destroying this species that pose a danger to humans, it produces dirty streams and swamps in which mosquitos reproduce uncontrollably. Pesticides do not work, and mosquitoes become a great threat to humanity with their deadly diseases.

Science and technology are not as perfect as nature, but they form the most advanced ecosystem that human beings have developed over centuries. Every new scientific finding, every new technology developed is the input and source of information for the next and more advanced system. Like all beings in nature, every invention can have terrible consequences for humanity, to the extent that it could destroy the world when out of control. That is why nuclear, biological, and chemical facilities are privately protected. No-one, unless they are authorized, is even brought closer to these facilities.

Do open-source platforms turn into a deepfake swamp?

As the world goes online, the threats and dangers that target humanity become digital. We almost opened Pandora’s box with these AI technologies that eliminate the limits of human intelligence with the help of algorithms over artificial neural networks produced by modeling the human brain. It seems that deepfake software is no different from dangerous chemicals. Such a perception is imposed on the online community. However, it is clear that deepfake tools are easily accessible from open-source software libraries. This availability speeds up deepfake’s motion and mutation.

New York University (NYU) researchers reveal that free and downloadable open source software is the primary factor in the commodification of deepfake technology. Many of these projects are held and shared on Github. Many project creators even request user donations through Patreon, Paypal, or Bitcoin. Determining that the software is mostly related to face changing and synthetic voice production, the NYU research team emphasizes that most of these tools cater to deepfake professionals rather than amateur developers because they often require programming experience and a powerful graphics processor (GPU). However, some have pointed out that some popular open-source deepfake platforms have increased the speed of amateur developers using these tools, thanks to the published user guides and tutorials of discussion groups.

Microsoft bought GitHub for $7.5 billion in 2018

The research titled “Deepfakes: Uncovering hardcore open sources on GitHub,” published in October 2019 with the signature of Rachael Winter and Anastasia Salter, also highlights the active role of GitHub in the process. The study pointed out that the user who first introduced the concept of “deepfakes” on Reddit.com and whose identity is still unknown today, the first pornographic deepfake samples published in late 2017 have been banned gradually since the beginning of 2018. However, it is argued that the main project is carried out with a series of open-source software shared on the GitHub platform. The study examines the platform role that GitHub plays in spreading deepfakes and questions open-source ethics. As a result, it concluded that the deepfake project has been deliberately spread to areas with limited moderation, with the support of open-source code, thereby maintaining its existence and development. The article titled “Analysis of the commodification of Deepfakes,” publish ed in the NYU Legislation and Public Policy Journal in February by Robert Volkert and Henry Ajder, points to the contribution of open-source platforms. The article states that open-source platforms are the starting point for open commercialization of deepfakes and the increasing speed of circulation. It is argued that deepfake services and market places that have begun to emerge also improve the commodification process.

As deepfake examples and popularity increase, the demand for a better code that creates and enhances such content also increases. Many important tools, such as Faceswap, DeepFaceLab, and Avatarify, which are the cornerstones of the Deepfakes development process, are published in GitHub libraries. Open-source platforms, such as Github or TensorFlow, continue to make significant contributions to developers in developing algorithms and creating easy-to-use mobile applications, despite risks such as potential vulnerability in codes and license incompatibility.

GitHub is an open-source platform that the software world knows very well and applies most. Many companies and software developers use the tools published on GitHub on their desktop, cloud-based software, or mobile application development processes. It was announced on June 4, 2018, that GitHub was sold to technology giant Microsoft for $7.5 billion. Since then, Microsoft has owned the largest platform of open-source software.

The result is a 65% performance in DFDC. There is hope, but no solution yet.

The results of the Deepfake Detection Challenge (DFDC) supported by Facebook, Microsoft, and Amazon could only be announced on June 12 because of the pandemic. Facebook announced on ai.facebook.com that the best performing deepfake detection model in the first stage of the competition, which was launched with a data set of 115 thousand videos on December 11, achieved an average sensitivity of 82.56%. The model, which completed the final stage in the first place on the previously unpublished closed-circuit data set and was the first in the competition, was able to reach an average performance of only 65.18%. We will evaluate (in another article) the results of the first global deepfake detection model struggle in which 2114 participants competed with more than 35 thousand models. However, the important result is that only two-thirds of the road has been completed for 100% detection of deepfake. But deepfake technologies do not stop and wait. New deepfake tools are constantly being dev eloped. That is why Facebook has decided to share the data set in the final stage, which was created only in 38.5 days with more than 3500 actors, with all the developers. Facebook has announced that it will publish this dataset with open-source code as raw data in order to contribute to the development of new generation AI-based advanced deepfake detection models.

It is hard to close the gap without reforming open-source platforms

In the statement in Facebook’s announcement, DFDC demonstrated the importance of “learning to generalize to unpredictable deepfake examples.” This means that in order to achieve 100% performance of deepfake detection models, achieving dominance over known deepfake technologies is not enough. It is also necessary to design AI models that have intuitive sensitivity to newly developed and not yet developed deepfake tools.

So, those who develop deepfake detection models are expected to be faster than those who developed deepfake tools to gain an edge over the algorithm war before it is too late. For this, don’t they deserve some time and positive discrimination? Isn’t it necessary to tidy up the open-source platforms that provide unconditional and unregulated ammunition and speed up those who develop deepfake tools and produce deepfakes with the help of these that are not known by whom? There is no need for blocking or banning. Even if the sharing of open-source code of deepfake tools is tied to a record and rule, will the speed of the deepfake threat be slowed down? Is it more difficult to maintain the supreme values of personal rights, peace, security, and democracy than to run million-dollar competitions?

Do biometric videos and audio data recordings protect against deepfakes?

Cybersecurity initiatives to develop measures against and solutions for deepfakes are progressing through two channels. Under the leadership and with the support of the online world’s giants (e.g., Facebook and Google), R&D initiatives, on one hand, are developing detection models that can distinguish fake synthetic images and sounds from reality. On the other hand, security-based institutions are trying to create a target-oriented personalized cyber defense shield through biometric video and audio to eliminate corporate risks against deepfakes. At this point, the following question can be asked: It seems to take time to develop a general detection model that will provide 100% security against deepfakes. So, can the personalized biometric video and audio security be a deterrent solution for targets, such that the production of their deepfakes will cause great harm?

As the quality of the video becomes an authentication tool, the risk grows

In a previous article titled “Deepfake has also an eye on our money,” we mentioned that western banks, which have started to use a “Customer Video Authentication” system, assumed the risk of deepfakes. The International Financial Auditing Agencies have stretched the “Knowing Your Customer” (KYC) obligation to banks in the fight against “Fraud, Anti-Money Laundering, and Terrorist Financing” (AML). With the regulation of the Fourth Anti-Money Laundering Directives of the European Union, dated 20 May 2017, it paved the way for European Banks to identify its customers through “video calls.” Thus, the banks initiated the service for banking transactions through video calls on the condition that they scan their legal identification documents, without coming to the branch, via the web or mobile application. But this time, the main problem has become how to distinguish whether the person in the video is real or a realtime deepfake.

This concern (whether a person is real or not) is already the threat of deepfakes. The U.S. Army is preparing to bring a person-oriented solution to this problem and to resolve the suspicion of deepfakes. Within the scope of the 2020 Defense Budget signed by Trump, the President of the United States, the U.S. Congress, which was especially concerned about China and Russia, loosened its purse strings. An annual budget of $5 million was provided by the Intelligence Advanced Research Projects Activity (IARPA) affiliated with the National Intelligence Office as the prize in the competition to develop deepfake detection models. Researchers of the U.S. Defense Advanced Research Projects Agency (DARPA) are also working on technologies that aim to automatically detect the integrity of images or videos, whether they have been altered or not, within the scope of the MediFor Project. The MediFor Program aims to create an end-to-end media forensics platform that can detect manipulations and deta il how manipulations are done. The Pentagon did not find all these initiatives sufficient or believed that they would give results in the short term, and it has put in place person-oriented biometric deepfake measures simultaneously to at least secure its own structuring.

U.S. Army compares deepfake targets with biometric media database

At the Aberdeen Proving Ground War Ability Development Command in Maryland, researchers at the Science and Technology Directorate are making efforts to meet urgent operational demands regarding biometric security. The Intelligence Systems and Processing Department of the Directorate is working on two biometric systems to increase the security of the Armed Forces and Intelligence members on the field and protect them from deepfake traps. These two systems, Video Identity Collection Abuse Prevention (VICE) and Audio Biometric Identity Abuse Prevention System (VIBES), interact with authorized biometric identification platforms to detect media (e.g., photo, video, audio files). Biometric systems provide identity assurance by matching IDs with media stored in the database. This makes it safer to decide whether a person is empowered for a particular task or to access a system, or it is threatening.

10-year biometric media database

It is stated that both biometric systems, VICE and VIBES, contain 10-year data consisting of photographs, videos, audios, and other media. It is noted that the developed algorithms presented good results in matching with the media database. Thanks to the two biometric systems, it is possible to distinguish what is real from the deepfakes by matching the audio and video data. For example, in the video of an ISIS leader, it was determined that the sound did not match that person’s real voice identity. Undoubtedly, the video and audio biometric identification records of all those involved in the Armed Forces and Intelligence duties and powers are included in this strategic media database. Thus, many areas of national security are trying to avoid deepfake traps by matching the biometric audio and video identity, from the reality of the commander’s instructions to the security of the on-site staff, as well as the correct and effective determination of search-and-rescue targets.

The national security role of biometric video authentication in the United States is not limited to the Ministry of Defense. Tygart Technology, which is a server-based video and photo forensic analysis system and whose main product is MXSERVER, provides state and federal government customers in the United States with a biometric recognition infrastructure. The company contributes to security through image matching by processing video and photo collections into searchable sources.

Deepfake’s enigma in the biometric security market

The global biometric market is expected to increase by around $30 billion, with a growth of over 30% annually from 2018 to 2023. With the cloud-based subscription services security market, Pentagon’s demand for Advanced Authentication Technologies is expected to accelerate growth in the biometric security industry. The Australian Digital Transformation Agency (DTA) has also announced that it will integrate biometric authentication with myGov citizen services.

Many experts in the field of biometric security emphasize that the most important threat expected from 2020 is “synthetic identity.” It is stated that synthetic identity fraud, which is considered one of the most difficult types of fraud, will increase, and financial organizations will apply to the companies that develop technology in the field of biometric authentication to prevent this. Naturally, the organizations will be looking for a powerful and secure biometric screening system that cannot be easily deceived by deepfakes. Biometric security companies will have to invest in more sophisticated authentication technologies to guard against synthetic identities. For example, Ben Cunningham, Pindrop’s Marketing Director, working in the field of anti-counterfeiting technologies for voice biometric authentication and calling, shared his efforts to increase Pindrop’s activities against the deepfake threat at the SpeechTEK Conference in November 2019. Cunningham stated that while analyz ing the outputs of the voice calls coming to the call center for biometric authentication, they are now concentrating on sound data that the ear cannot hear. 

The claim that 95% of Deepfakes overcomes biometric security

The biometric security industry may be very reliant on biometric processes with precise mathematical models that meet high-security criteria. Through biometric technologies, they will come up with various theses, arguing that the input can be a real person or a computer-generated video. This is a billions-of-dollars market where the customers from the armed forces of the countries that invest in the most advanced technology to the international financial institutions make purchases with abundant zero budgets. On the other hand, there is significant research that is revealing that biometric security is incapable of detecting deepfakes.

In Switzerland, the Idiap Research Institute, a non-profit organization, researches speech processing, computer vision, biometric authentication, multimodal interaction, and machine learning. According to research done by the Institute of Pavel Korshunov and Sebastien Marcel, 95% of deepfakes easily overcome biometric facial recognition systems. According to Korshunov and Marcel, current facial recognition systems are vulnerable to high-quality fake videos created using Generative Adversarial Networks (GANs). Consequently, it is not enough to match the synthetic fake faces created by GAN to the database, but a model that will be detected automatically is needed. The researchers used open-source software to create deepfake videos with faces transformed with a GAN-based algorithm. They explained that the state-of-the-art facial recognition systems based on the VGG and Facenet neural networks are vulnerable to deepfake videos as a result of 95% false acceptance rates.

As a matter of fact, fighting deepfakes is a war of algorithms based on artificial intelligence (AI) between attackers and defenders, regardless of its method, model, and technology. And in this war, none of the parties can claim absolute superiority for now. Whoever has the upper hand instantly, the other party pays a heavy price.

Most of the AI and political and scientific articles focus on deadly autonomous weapon systems, such as “killer robots” that detect and destroy targets without human intervention. However, AI is widely used in a variety of national security missions. AI consists of an important part of the software used to operate physical systems such as autonomous aircraft or ships. As in the example of using machine learning to classify targets in satellite images, sometimes AI is considered a part of analytical processes. In either case, AI is not a military element on its own, although it has become an important power and basis for the efficiency of national security tasks and systems.

Is the war won on the field or in the operation center?

You have definitely seen war in action movies, serials, or war and terrorism on the news. A digital operations center is established for national security operations. Fights in the field with global terrorists or military elements of the enemy state are operated through these AI-aided bases with the most advanced and integrated digital image analysis possibilities. National security units in the field act according to the instructions of their superiors in the digital operations center. And, using AI-based algorithms, the operators in the operations center make strategic decisions by analyzing the image data obtained from digital sources in the sky, such as satellites or unmanned aerial vehicles. Image is a crucial factor of the intelligence. If the flow of images from the field is interrupted, the national security operation remains blind, deaf, and dumb, and the chance of success is reduced. What happens if the intelligence images are not real but synthetically produced, altered, o r manipulated? Wouldn’t the war in the field be lost?

Naturally the national security does not just consist of military operations beyond the borders of the country or outside the residential areas. Threats targeting social peace and comfort, which we describe as public order in a broad sense, are diverse. Terrorist actions in cities, social movements leading to anarchy, protest demonstrations, street conflicts, major traffic accidents, acts of God (i.e., floods, fire, earthquakes), and even public-order-oriented attacks (such as robbery, extortion, looting) pose a great danger to public safety. The public authority, which is responsible for the security of the cities and the country, monitors, inspects, and controls the cities, highways, and all strategic areas, especially the critical areas, with CCTV camera security systems. The ability of security forces, health care teams, and firefighters to respond to predicted or unexpected hazards in the fastest and most effective way depends on the video data flow obtained from security camera s. In the article titled “How safe are security cameras against deepfake?” we talked about the risk of the infiltration of manipulated synthetic video content into the CCTV security camera network. But the social threat of deepfakes is not limited to the CCTV security camera network in residential areas.

Can air defense be defended against deepfakes?

In the COVID-19 coronavirus pandemic process, quarantine, social distancing, and mask control in the city centers of many countries are carried out with the image flow obtained from drone clusters. Critical security decisions can be drained when the national security network in the sky that keeps life under surveillance turns into a source of disinformation that could mislead public authority. Hyper-realistic synthetic images can eliminate the possibility of performing correct and necessary intervention in a timely manner. Deepfakes, the biggest threat to the digital vision of national security in the sky, are likely to become the most dangerous secret enemy of nation-states in the near future. When it came to air defense in the past, the first thing that comes to mind would be the fleet power consisting of high-tech aircraft. Now, it is necessary to defend the air defense system against deepfakes.

Citadel Defense, offering AI-backed autonomous drone security technologies to military, political, commercial, and international customers against airborne espionage attempts and attacks, took an important step last month. The company, which serves the U.S. Special Forces, Air and Naval Forces, also provides systems to protect valuable assets, such as merchant marine fleets, oil wells and refineries, against airborne attacks. Citadel Defense is creating a new defense structure against identity fraud tactics. At the end of last month, the company announced the release of a new software that contained artificial neural networks in deepfake technologies as a strong defense against enemy fraud tactics. The company explained that the software is designed to fight increasing enemy tactics that tend to tamper with the security intelligence equipment. The solution, called “Adversarial Networks,” will be distributed to the Citadel’s Titan C-UAS drone defense system platform.

Signal classification model that separates the enemy drone from deepfakes.
Christopher Williams, the CEO of Citadel Defense, said that just as anti-virus programs have methods to detect software abuses, Titan has also automated methods that proactively defend against identity fraud abuses. Williams stated that they added new deep-learning skills to Titan for this purpose, and thus, they aim to blind the drone-equipped enemy. He also stated that they aim to prevent the enemy from gaining any advantage or secure action opportunity in this way in controversial and complex radiofrequency environments. Classification models that are developed in the system and use registered image creation algorithms measure the authenticity of a detected signal. They help to determine whether the signal is a real drone or a fake signal synthetically produced by the enemy.

The U.S.A. spends billions of dollars on the Cyber Forces Command.

The U.S.A. Armed Forces established the Cyber Command (ARCYBER) exactly ten years ago. Billions of dollars have been spent each year in ARCYBER’s cyber training and cyber drills that increase digital war power. The U.S.A. Armed Forces took another big step to move digital war and sky intelligence superiority beyond the atmosphere. With the approval of the 2020 Defense Budget, also known as the National Defense Authorization Act by the U.S.A. president Donald Trump, the U.S.A. Space Force Command has been officially established this year.
The progressing military use of AI is also changing the way military operations are carried out by states. AI-based technologies, autonomous drone clusters, and algorithms that quickly review large amounts of information will undoubtedly increase military fighting speed and efficiency. Research and development work on AI promise to provide sophisticated accuracy and efficiency to complex and dangerous tasks in the field of national security. However, solutions for potential risks that may arise in the context of national security or multinational military operations are not yet sufficient.

Data-poisoning risk makes AI’s reliability questionable.

As AI-backed war becomes widespread, military cooperation seems to be getting harder and harder. It may not be easy for states, having AI technology, to overcome political obstacles in sharing sensitive data required to develop and operate AI-aided systems with their allies. In AI-aided international security operations, the data shared by the allies can leave them vulnerable, especially against manipulation. They are concerned that the enemy can enter their own data repositories and “poison” the data, threatening their national security by adding fake data or deliberately making existing data defective, so they are not wrong at all.

Enemy forces can use AI to launch deception campaigns aimed at interfering with an alliance’s military command and control processes. They can also use AI to launch misinformation campaigns that will lead to disputes among allies. In this way, the enemy can mislead AI-based target recognition systems, lead the system to overlook military targets or to classify them as “not military targets.” It can poison image data to identify civil infrastructure as military facilities. The worst part is that it can cause the wrong targets to be hit. Those who plan and manage military operations have tried to deceive their enemies during times of war and crisis for a long time. In World War II, the allied forces applied to a complex deception that contained imaginary army images equipped with swelling tanks and aircraft traps to deceive Nazi planners about the location of the D-Day landings. There is no longer a need for such physical deception in military operations because AI offers a powerful pe rsuasive weapon (e.g., deepfakes) for setting up digital traps and disinformation.

Deepfakes is a dirty weapon in AI wars.

Deepfakes can be used in unexpected ways, as a dirty weapon that does not actually comply with war ethics. The enemy can generate the deepfake of senior commanders to give false or contradictory orders to the troops in the field. For this, they can use the video or audio recordings of a real commander, obtained from the media or through intelligence. Or, to create unreal satellite intelligence images, it can build hyper-realistic synthetic media through productive contended networks. These fake commands and intelligence reports transmitted via video teleconferencing, phone, email, or radio can cause troops to be deployed to assist the enemy.

Alliance military forces can be even more vulnerable against the disinformation through deepfakes because it is more difficult to ensure coordination in multinational command and control processes. The time pressure, stress factors, and complexity of the national security operations increase the probability of people being enlisted who are not experienced enough to perform deepfake commands without any question.

As the quality of deepfake production increases and it becomes harder to decode the reality of falsified content, national security problems will increase for all states. Not only our online cybersecurity but also our life and freedom will be in danger.

Deepfake Market Research

The research industry strategically contributes to the marketing and advertising industry. They are the main data source of strategic decisions and creative campaigns for marketing. With the data obtained from the research companies, the target audience is identified, the messages are determined, the language and brand face to be used are ascertained. Research companies determine consumer trends, changes in patterns of consumption, type of perception, the pulse of the market, and future trends. Research companies have started to get involved to guide the marketing sector on deepfake investments.

In the United States, Market Research Media, which conducts market research for the government, Congress, large corporate companies and investors, and provides data to nearly half of the giant companies in Fortune 500, published a deepfake research project for the marketing industry. In the summary of the report titled “Market Analysis of Deepfakes Technology,” issued at the end of February and published on Marketanalysis.com, the opportunities offered by deepfakes to the marketing world are listed under three headings: entertainment, e-commerce, and communication. Under the heading of entertainment, it is noted that with synthetic human voices and images that are indistinguishable from the original, convincing stories can be produced even if the creative effects and real characters do not exist in the movies. Under the heading of e-commerce, it is emphasized that deepfakes can provide consumers with an advanced personalized experience during the selection of products. Under the head ing of communication, attention is drawn to the opportunity to benefit from deepfake brand faces that can speak every language with voice synthesis and face manipulation. Of course, this also provides the opportunity for a meeting with the new markets. MarketAnalysis.com sells the entire research for $5,950 to a single user and $9,950 to global websites.

Inevitablehuman.com, the online human community platform of the future, offers us the opportunities listed in the Market Analysis of Deepfake Technology as well as other opportunities that will inspire the marketing industry—for example, special video campaigns for each demographic group, step-by-step video access for each user, and even the opportunity to prepare single-use video messages for each customer.

Catalog Sales in Deepfake

Brand administrators need to pay thousands of dollars to access market information for business opportunities offered by deepfakes. But for advertising videos, it is much cheaper to rent deepfake actors and brand faces. Moreover, as in 2019, it is not necessary to ask a deepfaker to produce the characters they designed by using deepfake technologies. We mentioned in a previous article (“Deepfake Gunmen shoot even $20”) that the deepfake customers, who we cannot know if they are well-intentioned or malignant, can find deepfakers at websites such as Freelancer.com and Workana.com. The price to produce a one-minute deepfake video start at $20 and range to $50 or $100. But there are no references other than the demo video to prove that the video was the freelance deepfaker’s own production. In addition, the requested time for production was around three months. However, there is no need to wait for the order anymore, as serial deepfake production has already started. You can see them on the internet and choose what you like.

Synthesia, the start-up that prepared the popular deepfake public service ad in which David Beckham spoke in nine different languages for a social responsibility campaign, is one of the pioneers of Catalog Deepfake Sales. The company was founded four years ago by a group of researchers and entrepreneurs from UCL, the Technical University of Munich, Stanford, and Cambridge. They offer localized and personalized videos on their website Synthesia.oi, deepfakes suitable for different languages and scenarios. The CEO of Synthesia, named Victor Riparbelli, argues that within three years, completely convincing digital people can be available for an interview with Business Insider.

Alethea.ai, which takes its name from the first female name derived in Greek, took into service the first blockchain-based deepfake marketplace, which will save customers the struggle of ordering a deepfake. Alethea.ai gathers the world of marketing and media with deepfakers that produce deepfake characters with AI-based algorithms. The photo-realistic digital personalities that can voice fictional texts with synthetic lip movements produced by contracted deepfakers are sold to anyone who pays its price through Alethea.ai. Purchasers of deepfake characters displaying different profiles with different appearances and physical features can add them to another original video or play them in their own videos according to their scenarios. Advertising agencies, production companies, or any person or institution that want to produce deepfake videos can choose profiles that are suitable for their scenario from the digital catalog on the website of Aletha.ai, according to their personalities, characteristics, and physical features. Deepfake faces try to affect their customers with gestures, not even voice. The cost of fake digital actors varies between $49 and $99, depending on the character, and it includes a 200-word voiceover.

Cryptocurrency genius deepfaker put up its deepfake character with its own identity for sale

Alethea.ai can also offer digital characters that are a copy of a real person to the deepfake market, alongside the catalog sales of deepfake characters designed by deepfakers. Deepfakers, who make a revenue partnership with Alethea.ai, can also put up their licensed digital copies for sale as well as the synthetic characters they produce by copying other real people. Alex Masmej is one of the deepfakers who aims to earn money from his licensed deepfake that he produced with his own identity and personality.

Twenty-three-year-old Alex Masmej is one of today’s digital geniuses. He developed his own cryptocurrency ALEX based on Ethereum, an alternative digital currency to Bitcoin. The young Frenchman is also leading the production and marketing of his own deepfake in cooperation with Alethea.ai. In Alethea.ai’s deepfakes catalog, a character, modeled by Alex Masmej, was presented to customers in the second half of last month. The rising star of the cryptocurrency market, which has beco me popular with its $20,000 share sales, does not hesitate to offer this deepfake to the market with its real-life identity and 200 words at $99. Masmej, who discloses in his interviews that he made this to increase his reputation, emphasizes that he has full control of his deepfake against misuse and malicious uses that could endanger his future.

Does Baby O (bama) deepfake create a smile or tension?

Brands, agencies, and production companies who prefer to use cloned deepfake characters of real people instead of designed avatars cannot do this in every scenario. Alethea.ai claims to build its activities on the concept of “Legal deepfake” and requires the consent and approval of the real person for their characters. The deepfake cannot be produced without the consent of the real person. Furthermore, the approval of the real person is required for the scenario or the speech of the deepfake character. Alethea.ai, which offers Baby O to its customers for $99, must have gained the approval of the former U.S.A. President Obama due to the similarity. Otherwise, those who try to buy Baby O and use it for malicious purposes can cause big problems, especially now that the United States is going through critical times after George Floyd was killed by racist policemen. If Baby O said the wrong words, along with one of the contracted deepfake producers named GAN Master, Alethea.ai could be in trouble.

In his interview on Cointelegraph on May 15th, the CEO of Alethea.ai, named Arif Khan, argued that he could provide a steady infrastructure to license, distribute responsibly, and earn money with legally permitted creations, as opposed to the unlabeled and potentially harmful deepfake content of blockchain. Khan noted that all content created for Alethea.ai’s marketplace is tagged as controllable in collaboration with the software firm Oasis Labs and said that secure blockchain verification can block suspicious content, similar to Twitter’s blue confirmation verification. Khan said that the legal permits and the consent of individuals will be key criteria for deepfakes that can be circulated and used to earn money.
It is not known whether Alethea.ai’s permitted and registered deepfake model can raise awareness of the positive use of AI-based synthetic media technologies, as its CEO claims. But in the United States, the world’s largest brand exporter and marketing giant, it is possible to say that among the 30 million unemployed people, many have the option to earn money as actors, especially those in Hollywood.

The social distancing rule, having entered into our lives with the COVID-19 pandemic, has disturbed the balance. Social isolation has taken over the world, businesses have closed, transportation has stopped, working at home has become widespread, but life goes on. Consumption also continues. So, we continue on to competition and marketing!

If we read this sentence backward, we can say that if marketing continues, life goes on. However, advertisers are sitting in the state room of marketing and have not been able to go to their agencies for weeks. They have tried to protect their brands from their computers at home. On the other hand, we know that whatever trails us in life serves marketing purposes without ever being noticed. Sport is not real competition, nor is art only a reflection of culture and aesthetics, nor is technology just a printout of science. Each of them delivers emotional messages to encourage us to adopt and purchase brands, products, and services. The economy of each one grows with the contribution it makes to the marketing company. Fashion is more transparent than them at least. It lives openly its close relationship with marketing.

The Internet: Master of the Marketing World

From the first months of 2020, due to the nightmare that the world has experienced, things have also come to a standstill in the marketing world. Is it easy to maintain the advertising and marketing cycle as before by wearing a mask and with social distancing? While advertising expenses exceeded the half-a-trillion-dollar limit in 2016, it has been estimated that it has been exceeded by $560 billion last year. We will see that the COVID-19 pandemic will have made “a journey to the past” in the marketing world this year. But we know that digital advertising takes the lion’s share spending $333 billion on advertising, with TV commercials following it with $188 billion. These two are almost $560 billion. Global social networks such as Facebook, Twitter, Youtube, and Instagram are the biggest slices of the digital advertising cake. While it is estimated that digital advertising will exceed half-a-trillion dollars alone and grow by nearly 100% in 2023, a decrease even in TV ads is expecte d. This scene shows us two things about marketing and advertising. First, the power of visual communication ensures that video advertising content is increasingly included in the marketing world. Second, the internet is completely superior to TV in this field and looks like it will progress to be the sole advertising media in the near future

Advertisers Could Not Cope with the Pandemic

The COVID-19 pandemic has interrupted the growth of marketing at just this point. The advertising world cannot go outside, and current video marketing contents cannot be shot. Each of the current global advertisers, such as Samsung, Procter & Gamble, L’Oréal, Unilever, Volkswagen, has been spending approximately $10 million annually for advertising. However, for their ordinary big-production commercials, flashy shooting sets have not been set up for a while due to the pandemic. And, except for the Hollywood, Premier League, and NBA idols who have become the face of multiple brands, the advertising crews such as the gaffers, soundmen, cameramen, and make-up artists cannot meet on a set. Desk-bound advertising and trying to save the date by adding new texts and sounds to the stocked video is not enough for not only the consumers but also the brands.

Will Deepfakes Become a Lifesaver for Brands?

Are the advertising and marketing sectors already trying to manage their lifestyles and consumption styles by making people believe in fictional facts or not? So, who else but they go to work with deepfake technologies? It seems that the meeting of deepfakes and the advertising sector, which is one of the grey fields for turning it into a digital weapon, will pick up speed with the driving force of the pandemic.
A current example for focusing the creativity of advertising agencies on deepfakes in the pandemic process came from the United States. State Farm, one of the largest insurance groups in the country, transferred the lip movements of Kenny Mayne, a 60-year-old famous sports commentator on the channel, to a video that had been shot when he was younger. It was published in ESPN. Deepfake commercial films have been started to be published in the documentary named “Last Dance” about Michael Jordan and Chicago Bulls in ESPN at the end of April. In the commercial, Mayne announces the 6th Championship of the Bulls as if he is 38 years old. He prophesies what would happen and emphasizes that the success of Jordan and the Bulls is the kind that will lead ESPN to make a documentary about them. Mayne says, “It will be called something like Last Dance and will consist of a 10-episode series and be published in 2020.” The former logo of the famous insurance company appears in the background. The 2 8-years younger face of the famous presenter says, “And this clip will be published as a State Farm commercial film in this documentary.”

Deepfake Advertising Campaigns Started Last Year

Deepfakes had already started being used by big brands before the pandemic. In 2019, David Beckham, a British football idol, gave a message in nine languages that he did not know in a commercial for the charity called Malaria No More thanks to deepfake technology. In January 2019, Doritos, one of the brands of Pepsico, started a campaign with an app named Sway for chips lovers that enables them to make a video by using AI technologies. Doritos’s campaign was about the commercial film of Superbowl (that is, the National American Football League’s final), which drew more than 800 million viewers around the world. Doritos asked the campaigners to repeat the dance movements acted by famous rapper Lil Nas X and actor Sam Elliot and share the video using the Sway app.

Deepfakes also provide the marketing world with the opportunity to offer its customers more advanced personalization experiences. Targent.ai has developed an algorithm, based on the fact that the purchase intention of the consumers are high when they have personalized experiences. For example, if a female customer is able to customize the skin, hair, lipstick color, and even the ethnicity and race of the model that exhibits the product, it becomes even easier to sell the product to her.
It is understood that even when there is not a pandemic, the advertising industry is no longer interested in spending millions of dollars on commercial film sets, actors, and a crew. Deepfakes offer them the opportunity to show their creativity by sitting at the computer, brainwashing people with the messages of the brands. The single issue is how and with whom will the advertising agencies and production companies produce the deepfake synthetic images and sounds that they designed in the development process of deepfake technologies?

Let the geniuses of fictional reality not worry about it. We will mention the “deepfake markets,” in the next article.

Protocols ensure quality standards, but how about the security standard?

Total Quality Management used to be one of the most coveted development areas, despite its waning significance for corporations. It was common for companies to compete with each other while going through long and arduous efforts with accreditation companies to amass new ISO quality certifications. Quality was measured by the highest level of benefit from the product or service for the customers or the end-users. Consumers perceived quality based on objective criteria, such as the most durable, most healthy, most tasty, or most developed functions depending on the product or service type. ISO stands for the International Organization for Standardization. It is an organization that establishes quality standards and has 164 members, each from a different country. The ISO measures value that is not directly the quality of the end product or service. It aims to establish quality by instituting a standard on the operation and production processes. What makes this possible are the written p rotocols that establish a standard on business systems.

Security protocols are vital for all

Protocols on areas of security are vital. There are protocols in place for any emergency. In the United States, you would call 911 in an emergency or 112 in another country. Work safety protocols that the employees must follow at an iron and steel plant or an oil refinery differ from those at a construction site. Different protocols are enforced for each emergency situation that may arise at a health or education institution. When security is at stake, awareness and sensitivity for protocols become greater. Compliance with protocols by those in charge sets the tone for the security of themselves and others.

The dawn of the digital era brought on a swift transformation in the conditions of online security. Today, life is no longer just about the physical entity of the earth we live in. When we get up in the morning, we probably say good morning to the people on our social network before those in our homes. The person we confide may not be someone close to us physically but someone on another continent. We hold job interviews, not in the meeting rooms of the companies we hope to work at, but on the work desk of our home by connecting to the HR manager on a program like Zoom, Google Meeting, Skype, or WhatsApp at an appointed time. We click on the internet banking application on our computer or mobile phone, rather than going to the bank branch just a couple blocks away. Consequently, the quality of our online life depends on cyber-security.

Social engineering has not played its deepfakes trump card yet

Digital transformation and online social lifestyle are areas of social engineering. Wikipedia explains, “Social engineering is the effort to scheme people into providing certain information by taking advantage of their weaknesses through various persuasion and deception methods. It involves techniques aimed at changing people’s decision-making process.” So, the cyber-hackers dictate your lifestyle, your online value judgments, behavioral patterns, and reactions. Usually, multiple cyber methods are utilized simultaneously, once the determination has been made on the means and techniques to dupe you in the fastest, easiest, and most harmful way. Your security vulnerability and weakness are exploited before you even get a chance to know what is going on. Deepfakes are the next step in the cyber-attacks that are driven by social engineering, according to Rod Soto, an independent cybersecurity expert. That is why you need to have an anti-deepfake protocol in place from today, along with o ther cybersecurity measures.

Voice cloning technologies herald the onset of deepfakes

Zero-day attacks come on the day when security vulnerabilities are discovered. Thus, you have no time to alleviate the weaknesses. But, deepfakes are striking with ferocity. Fast and eerie developments, especially in voice cloning technology, are threatening the authenticity of all calls made on digital devices through any communication technology, particularly GSM. For now, only prominent personalities are chosen as victims, probably because of the high cost of synthesizing and manipulating voices for malicious intentions. Rod Soto, a cyber-security researcher, explained to CNN Business that crime syndicates find investors to fund their schemes for big loots from wealthy victims. Traffic or airport noises are used in the background to make the cloned voice sound more convincing. From the way it looks, future victims will not necessarily have to be prominent politicians or company CEOs. As deepfake technologies advance with AI-based deep learning algorithms, it will not be possible t o repel social engineering attacks with a sharp mind or trained eyes and ears alone.

Which authentication tools to use on your anti-deepfake protocol?

The social engineering aspect of the deepfake threat is similar for online corporate and individual users. Social engineering attacks on corporate targets are implemented by manipulating individual online users who work as an employee or executive at the company. This requires individual online reflexes to be improved by an anti-deepfake protocol. We take measures beforehand for the older or pregnant family members even if they are in good shape. Similarly, we need to be on high alert against the deepfakes threat.

The phone call may appear to be from the person you know. It may be that very person we speak to and see on the audio or video call. That person may even send the same message in writing from a personal or corporate email account. Still, the first step of the security authentication must include being constantly aware and skeptical about the fact that the person you are talking to may not be who you think he is.

As deepfake technologies advance and become harder to detect, it will be necessary to implement anti-deepfake protocols that contain multiple controls and authentications. This means multi-layer security “scanning” on multiple devices to verify whether or not the audio and images received from digital tools and online channels are synthetic and fake.

Deepware Scanner: A candidate to become the most important authentication tool for your anti-deepfake protocol

Available on the deepware.ai website by Zemana for online individual and corporate users, Deepware Scanner provides AI-based technological support to scan received digital audio and images for deepfakes. You can scan for deepfakes by adding the link for the digital audio or image content on the deepware.ai website or by uploading the content on the system as a file, making this process the fundamental item of your anti-deepfake protocol for all your content. You could even integrate the Deepware Scanner into your corporate system by taking advantage of the fast-advancing authentication performance, which is at 93%.

Multi-Factor Identity Authentication against real-time attacks

Anti-deepfake protocols must be backed with Multi-Factor Identity Authentication (MFA) applications, especially for organizations and companies. Any combination of two or more factors is used as protection against fraudulent access. For instance, as is common practice in banking applications, a password is used along with an SMS code sent to the mobile device. This serves as a two-factor identity authentication performed online for access and operation authorization. When one of the factors is compromised, the system still remains secure. Anti-deepfake protocols developed against real-time deepfakes threats could contain updated security questions and passwords or multiple identity authentication methods such as is the case in multiple meeting sessions. As a way to authenticate that the person you are speaking to is really the person you think he is, the person in question may be asked to quote a secret password while also answering a pre-determined security question. This could be a measure to resort to rather than simply complying with the risky requests and instructions received during the meeting. Moreover, terminating the meeting and calling back to verify the identity could be one of the authentication methods to be applied.

While all these measures may seem over the top, the threat of deepfakes is real. We may have a chance to elude deepfake-based social engineering attacks without sustaining any harm if we gain cybersecurity awareness with no shortfalls or deficiencies and employ meticulously developed anti-deepfake protocols and advanced security tools.

The world is no longer a safe place. We need to reshape the flow of daily life by staying alert against any unexpected threats and by taking measures to prevent us from getting fooled easily. Naturally, this is not a paranoiac or asocial detachment from life. On the contrary, it is adapting safe socializing as a way of life. So, don’t delay and establish your anti-deepfake protocol today.

That is the frightful and perilous story of the foundation of the United States, which is regarded as the cradle of technology. Threats driven by the supreme civilization of the era moved towards the unspoiled lands from the Atlantic Coast to the Midwest and West. California’s Silicon Valley on the West Coast of the United States is the destination of this treacherous journey. Globalization will never stop spreading Wild Capitalism to other continents by using the technology weapon. The method is the same in the Wild West and Wild Capitalism alike. The one who holds the weapon uses it to his own advantage. Civilization develops the weapon first, then abuses its power for its brutal objectives. However, it does not do it openly. It always seeks to achieve its objectives secretly and through deception. The power elites are a global extension of this mighty adventure as shown in the example of Operation Gladio, a “deep state organization” that, according to recent claims, was unearthed in Italy.

Technology develops weapons, but it is up to you to protect yourself

The depth of the danger to humanity poses a great threat. And now, this depth is completing its evolution of digitalization with a technology developed by using Artificial Intelligence (AI). Consequently, the cycle of using all of the earth’s resources and reserves for one’s own global interests has reached a new phase. Synthetic media that is created with deepfake sound and images have added a new dimension to these depths. The power to divide and conquer societies exists on a greater scale and has gained more speed. Just as in the Wild West, there is no-one to protect us from the perils of deepfakes. It is not known when, how, or from where the assault will come. “You can preserve your perception of reality only with your intelligence and mind. You must protect yourself with the technology weapons you can muster,” we are warned by Globalization.

World Economic Forum (WEF), one of the most important platforms from which current globalization messages are conveyed to the public, discussed deepfakes at the Davos Summit in February. As we briefly mentioned in a recent article, the Davos Summit hosted Hao Li, Professor of Computer Science at the University of Southern California, who is described as the world’s best deepfake developer by the MIT Technology Review. Pinscreen founder and CEO Hao Li and a team from the company demonstrated how participants could experience real-time face swaps on a Virtual Mirror developed by them. But the most important take away the clear message they conveyed: “You are responsible for your own security against deepfakes.”

Cheap fakes herald the dawn of the deepfake era

Today, as an indispensable part of our lives, digital media has grown into an incurable addiction for modern society. More than 1.2 trillion digital images and videos are captured annually in the world. This figure is increasing by 10% each year with more than 2.7 billion people capturing 85% of the images on smartphones. Increasingly, these digital recordings are being manipulated, and there are a higher number of digital processes on these digital audios, photos, and videos. Possibilities in audio and visual manipulation have moved beyond the expectations of experts, thanks to the image and audio processing technologies that are developed with deep learning algorithms through AI-based artificial neural networks. The confidence in audio and visual truth has gradually waned with the speed of growth and the spread of deep synthetic media. Generative Adversarial Networks (GAN), which demand high processing and graphic card capacity, and the other machine learning-based deepfake technol ogies still have not reached a high enough production capacity to occupy the public domain. But, cheap fakes—entertaining face swap apps on smartphones—or more elementary methods, such as speeding up and slowing down for the purpose of disinformation and online fraud, are used at a sufficient level to cause uproar and concern. The number of productions is rising also in deepvoice and deepfake media, which are becoming increasingly hard to distinguish from the originals. This will inevitably result in a “break from reality” that is sure to disrupt the processes of perception and decision-making for businesses and society.

Will it ever be possible to detect deepfakes in real time?

Truepic does not offer users the absolute scanning experience that Zemana Deepware Scanner offers. Nevertheless, it describes itself as a photo and video verifying platform that offers services to journalists, NGOs, public institutions, and companies. “It will never be possible to detect and stop deepfake synthetic media in real time at the root,” asserted Mounir Ibrahim, Truepic’s Vice President, and he also made the aforementioned points in an article on the WEF website about a month after the Davos Summit. There are a number of reasons for this, one of which is that as digital media content goes around the internet, almost all of the metadata, the essential information that defines it, gets lost, destroyed or changed.

Upon scanning the detection system of an image, it will be impossible to recreate the lost metadata. Thus, details such as the original date, time, and location of the image will probably remain unidentified. Also, nearly all digital images are instantly compressed a nd resized immediately after being shared on the internet. At times, this may be an effort to mislead the content viewer. Once they are uploaded or transferred, recompression and resizing of the images make it very hard to detect any pixel manipulation. Even if an automatic or machine learning-based detection model is registered and becomes available for use, cyber-hackers would develop other methods to remain undetectable.

The source of the outbreak: social media

What makes it even harder to detect fake synthetic media is the social media that spreads any content, regardless of whether it is real or fake, within seconds throughout the world. People with malicious intent can instantly plug fake content on social media platforms. Therefore, it may be too late to halt the spread of fake content even with a proven scanning model. Lack of awareness or fanaticism will speed up the viral spread of disinformation and make objective decision-making burdensome.

At the World Economic Forum, Truepic’s Vice President claimed that it would never be possible to detect and stop fake synthetic content before it spreads throughout the online world. The solution he puts forward is to establish their authenticity based on provenance (source, root, origin), instead of trying to decide whether digital content is fake or not. Are these remarks a premarketing attempt for the new technology that the company seeks to develop? Or, despite all of the aforementioned hurdles, is it possible to quickly detect and stop fake synthetic audio and images at the root before they spread? Time will tell.

The results of the Deepfake Detection Challenge (DFDC) are in, with the expectation that a deepfake detection model will be developed for the online world. The outcome has not been officially confirmed by the organizers, and there has been some delay with the announced dates, most likely due to the pandemic. Next is the deepvoice competition initiated by Google. Sincere or not, the rulers of the online world seem to be looking for a technological solution to the deepfake nightmare that they have created. However, even if this possible, it will take a long time.

Deep civilization is now taking over people’s minds with this technology weapon. As it was in the Wild West, it is still up to us to protect ourselves against vicious attacks.

Technology in the Pandemic Vernacular

Disinformation can be just as fatal as an infection—as was evident when 30 Turkmen citizens died around the same time in Istanbul when they drank absolute alcohol based on the false news that “it protects from the coronavirus.” The spread of false information and news with malicious intent drag people into the swamp of disinformation. There is an avalanche of absurd posts from people who believe everything they see or hear. Misinformation, albeit inadvertently, spreads at a viral speed, causing results as catastrophic as disinformation. A mutated and uncontrollable virus is causing death on a global scale, while havoc runs through the online world, brought on by a virally disseminated infodemic with all kinds of disinformation and misinformation. Two eminent UN organizations, the United Nations Educational, Scientific and Cultural Organization (UNESCO) and the United Nations Development Programme are trying to create awareness against phony information on COVID-19; however, their efforts fall short in the face of the threat.

The most destructive wave of global lies is deepfakes, which create, alter, or transform digital content to appear real by using AI technologies that run on artificial neural networks and deep learning. Deepfakes pose a number of threats, ranging from cyber fraud to manipulation of democracy to local and global stability, and they have evolved into a new version in the era of the infodemic. The first version of Deepfakes 1.0 was put to use for fun through face swap apps in 2017 and 2018. At first, porn-related synthetic videos were produced by swapping faces of famous women, followed by parody videos that mocked politicians and other well-known personalities.

The era of Deepfakes 2.0?

Having been on the scene since last year, Deepfakes 2.0 boasts of far more plausible and profound effects. The infodemic period appears likely to provide momentum in the new era when the vitality of truth will be under attack. The eradication of society’s common ground on authenticity spells disaster for the shared future as well. For centuries, manipulated information obtained through chicanery has been used for wicked purposes. Today, technology allows it to happen at an unprecedented speed and scale. With the advancement of technology, it is not too long before a teenager will be able to create and disseminate a high-quality deepfake video on a cellphone from the comfort of his couch. “We don’t know where this is going to end,” said Matthew Turek, the program manager at the Defense Advanced Research Projects Agency (DARPA). “It may end up in an environment of zero trust.”

Heralded by Deepfakes 2.0, the new era is not confined to deepfake videos anymore. The advancement in voice cloning technologies has also created the onset of a deep voice. Deepfake and deep voice have assumed the form of a great threat, together and separately. 

Infodemic Deepfakes Scenarios

Deepfakes have an immense influence on an infodemic. A video and/or sound recording that is manipulated to disseminate disinformation could cause harm greater than even the culprit can envisage, thanks to the wave of misinformation by the unhinged masses who share it without bothering to question its authenticity. The variety of infodemic deepfakes scenarios are limited only by one’s will and imagination, as well as technology.

As we have said in previous articles, when companies shifted businesses from offices to homes, complying with social distancing rule, they lost sight of cybersecurity. Once a risk only for global companies with offices in different countries, the threat of deepfakes are now valid for companies of any scale. We know that fraudulent financial instructions can be created, setting companies up to lose millions of dollars by cloning the voice of a company executive or by forging a convincing real-time video conference call. During the pandemic, vulnerable smaller companies, as well as their employees, may fall victim to many vicious schemes, such as bogus fund drives, fake tax collection, or forged liens.

The pandemic is also an extraordinary period that tests the competency of politics and bureaucracy. In these challenging times, deepfake attacks could spell big trouble at home and internationally for prominent political leaders in the government or opposition. All eyes and ears are on them for a solution to get through these testing times. Political leaders will find it extremely difficult to recover their losses and regain their reputation if a fake synthetic voice or video recordings of them turn into an infodemic, sending the public a false message supposedly from them.

Real threat to world peace and social order

The COVID-19 pandemic comes at a time when biological and chemical proliferation has turned into a race, and cultural differences are manipulated by xenophobic discourse. The fight that each country is giving against the COVID-19 novel coronavirus today may evolve into enmity or even a new cold war between the countries that accuse or hold each other responsible once the pandemic has been curbed. As COVID-19 emerges and spreads through countries, cooperation and solidarity issues may arise, and a viral deepfake infodemic could end up sabotaging world peace because of the rage that may run rampant in the countries that pay a higher price. International diplomacy may have its work cut out for it in the face of voice and video recordings made with AI-based deepfakes technologies.

The threat of an infodemic on every country’s social peace and public order is as real as it is on international peace. Infodemic deepfake media content containing cloned voices or reproduced synthetic images could spread via social networks, provoking irreparable tragedies including panic, public unrest, violence, looting, and racist attacks in many countries.

An infodemic renders human life and health systems defenseless

An infodemic that escalates throughout the world as a combination of disinformation and misinformation would not only endanger people’s health and lives through false information and trajectories, but it would also present a great risk to public health by compromising the local health system. There would be endless dire consequences brought on by AI-based digital fake sound and video recordings that present protection methods against the coronavirus or self-treatment remedies during the quarantine period, supposedly by a medical doctor, or someone representing a professional health organization or a ministry, or an international health organization such as WHO. The pandemic is already overwhelming local health organizations’ bed and intensive care unit capacities, and it would create a greater burden with an infodemic assault precipitated by deepfakes messages that encourage the use of harmful chemicals, absolute alcohol, or disinfectants, or medicines that are developed for other illnesses against COVID-19.

The threat of an infodemic seems here to stay for an unforeseeable future, regardless of the upcoming deepfake detection models discovered as a result of the Deepfake Detection Challenge (DFDC) lead by Facebook, or the Deepfakes Accountability Act discussed in the United States, or the deepfake constrictions that global social networks such as Facebook and Twitter are trying to endorse. Once detected, infodemic content may be removed from social networks and websites; however, that does not stop them from turning into “zombie” content and finding their way into the internet archives. In fact, they may attract more attention than real and authentic content. Let’s cover zombie contents another day and remember, “A lie could prowl half-way around the world before the truth gets out.”

A security issue was bound to ensue once the number of Zoom users increased twenty-fold in three months. Zoombombing attacks, or Zoom raids, by uninvited guests can easily infiltrate video conferences or training sessions on Zoom. Hacking these sessions with inappropriate images and words has instantly caused a furor. It has been at the top of the agenda as if it was the only thing to worry about. In the United States, video conference hacking is now considered a “federal crime,” following a security statement made by the New York attorney general’s office, FBI’s Boston bureau, and Eric Yuan, Zoom’s CEO, in early April. In response to these rising Zoombombing attacks, the United States Attorney for Eastern Michigan issued a written statement on April 3, which asserted, “Disrupting a public meeting, invasion of a computer, using a computer to commit a crime, committing a hate crime, committing fraud or making a threat in this way may be punishable by fines and imprisonment.

US attorney may be trying to elude to deepfake

Matthew Schneider, United States Attorney for Eastern Michigan, said in a written statement that “If you interfere with a teleconference or public meeting in Michigan, you could have federal, state, or local law enforcement knocking at your door. You think Zoom bombing is funny? Let’s see how funny it is after you get arrested.” While discouraging Zoom hackers with the long arm of the law, he also expanded the definition of possible risks. Zoom hackers could hack not only into Zoom sessions but also into the participants’ cameras, compromising their private photos and secret and personal information. However, he did not even consider what Zoom hackers can do under the deepfake mask.

What if it wasn’t the real Boris Johnson on the Zoom screen?

Before he was diagnosed with the coronavirus and put in intensive care in the UK, Prime Minister Boris Johnson had a COVID-19 meeting on Zoom during the first days of self-quarantine at his home. Among the many possible attendees were ministers, high-ranking bureaucrats, and scientists. The health of the community and the future of one of Europe’s strongest nations were being discussed at the highest level, and it is possible that they were making decisions not only about the United Kingdom but also about the world. The UK spearheaded the “herd immunity” exercise in the world, giving the novel coronavirus free rein. In the end, when the prime minister’s COVID-19 test turned out positive, they had to make a U-turn and impose the quarantine measures. So, during those days, instead of the prime minister of the UK, what if it were a hacker with a Boris Johnson deepfake mask, imitating his voice on the Zoom screen? What if the hacker told the ministers and bureaucrats that the quarantine measures were not necessary? When Boris Johnson himself was performing so poorly with COVID-19, what would his impersonator mimicking his face and sound cost the empire where supposedly “the Sun never sets”?

Masquerade ball precedes the invention of deepfake

On March 11, a Paris court sentenced Gilbert Chikli, one of the Franco-Israeli defendants, to eleven years in jail and a $2-million fine. Meanwhile, Anthony Lazarevitsch, his partner in crime, received seven years of prison time and a $1-million fine. From 2015 to 2016, these two men dumbfounded everyone by simply using silicone masks to do what deepfake can do today. During the two years of this masquerade, 54-year old Gilbert Chikli and 35-year old Anthony Lazarevitch managed to trick people by constructing a silicone mask of Jean-Yves Le Drian, the former defense minister and current foreign minister of France. The duo swindled $55 million from prominent people in their criminal enterprise. The Guardian reported that these fraudsters targeted more than 150 wealthy businesspeople, politicians, and religious figures whom they duped with a video conference invitation made over the phone or Skype.

Ringleader Chikli put on a silicone mask to impersonate the French minister Le Drian wit h a duplicate of the minister’s office appearing as a backdrop. Crooks scammed their wealthy victims under the ruse that they needed financial aid to save the captives of religious terrorist organizations in the Middle East, especially in Syria. They made the biggest heist in 2016 when Karim Aga Han made five separate deposits of $20 million into the Polish and Chinese account numbers they had provided. Though three of the transfers stopped at the last moment, $7.7 million was never recovered. A couple of months later, it was reported that Turkish businessman İnan Kıraç was also tricked into sending $47 to the grifters as ransom money for two journalists kept hostage in Syria. Their third victim was Corinne Mentzelopoulos, a French-Greek businesswoman, known for her world-famous Port wine and property investments. She lost $5 million. Some of their other unsuccessful attempts included Gabon’s president Ali Bongo, the French AIDS charity Sidaction, chief executive of the cement co mpany Lafarge, and the archbishop of Paris. Chikli tried the same scam for the second time 10 years apart. He received a seven-year prison sentence in 2015 for similar racketeering attempts between 2005 and 2006. Chikli’s scam stories inspired the 2015 French movie Je Compte Sur Vous (I’m Counting on You). 

Does deepfake pose a threat to COVID-19 fund drives?

Chikli was caught, along with his accomplice Lazarevitsch, in Ukraine. When they were arrested, they had silicone mask photos of Prince Albert of Monaco, whom they had not yet had time to mimic. There is something far more convincing than a silicon mask: a deepfake synthetic media that is created with machine-learning algorithms using artificial intelligence (AI) technology and artificial neural networks. So, with the dawn of deepfake, it may be that the era of silicon masks is now over. Hackers no longer require silicon masks of a minister, prime minister, or president to scam people. This comes at a time when the television reports that leaders, such as Boris Johnson, are holding all of their meetings on Zoom or other video conference apps, and that countries and international organizations around the world are launching massive fund drives for the COVID-19 pandemic. How many people would actually dare to double-check the authenticity of a political leader who appears on their comp uter or telephone screens? You decide.

Is it that easy? Take a look! 

Up until now, you may have thought of deepfake as fake videos that are created to be indistinguishable from the original by using special algorithms with the help of artificial neural networks and deep learning. In that case, you may also believe they only consist of videos prepared in advance. However, AI technology makes live deepfakes possible. So, don’t be so sure that the image on the screen is the person you think you know when you answer a video call on apps such as Zoom, Whatsapp, Skype, or Facetime, all of which have that feature. It very well may be that there is someone malicious behind the image or sound, pretending to be someone you know to set a dreadful trap for you.

Hao Li, the associate professor of computer science at the University of Southern California, is internationally known for his research on deepfake. The MIT Technology Review described him as the “world’s best deepfake artist.” Recently, he put on an impressive show with a team from Pinscreen, the company he founded and now leads as its CEO. Hao Li and the Pinscreen team demonstrated real-time, interactive face-swapping during the World Economic Forum (WEF) in Davos, Switzerland, in January 2020. They made the attendees appear as celebrities such as Michelle Obama, Lionel Messi, Leonardo DiCaprio, and Will Smith on a live and real-time “Deep Mirror” on special screens. Pinscreen’s research aimed to advance “new deep learning” and “data-driven geometry processing algorithms.” The technology allows users to swap faces with those of famous people on a virtual mirror, heralding a step further for deepfake. Classic deepfake technologies require data. In contrast, the new beta zone technol ogy by Hao Li and his Pinscreen team works instantly, swapping a person’s face with that of another without needing to be trained. Face-swapping happens in real time.

Virtual deepfake cameras coming soon

Do you think this technology is still some time away from appearing on Zoom? Take a look at seemingly innocent and entertaining Aphrodite.ai. This application lets you create a deepfake in real time by using your computer’s web camera. At the moment, the beta version works only on desktop systems that use the latest versions of Chrome and Firefox. And, the development process is in the works. The Aphrodite.ai team is said to be working on a standalone application that allows the transformed video to be used as a virtual camera on Skype or other video conference applications.

Don’t Zoom without zooming in on the participants first

Make sure to err on the side of caution with online security while using Zoom and other video conference apps. It will not be too long before we find out if an effective deepfake detection model will emerge as a result of Facebook’s Deepfake Detection Challenge (DFDC). In any case, we must keep hope alive and continue to protect ourselves, just as we do against the coronavirus pandemic. Don’t use applications, such as Zoom, that do not offer end-to-end encryption just because you want to jump on the bandwagon. In order to increase security during your Zoom meetings, use a temporary user ID each time and activate the Waiting Room feature to see who is attempting to join the meeting. Also, go to Settings to disable the “Join Before Host” and other risky features. Finally, once the invited attendees join and the meeting begins, lock the meeting to outsiders. 

It is like a madhouse out there because of COVID-19, and what may be lurking around the corner is anyone’s guess. You may receive an unexpected video call from someone you cannot say no to. It is better to be safe than sorry. Keep vigilant!