Even though they tend to be similar, intriguing science fiction scenarios still draw audiences into the plot: The world is on the brink of a catastrophe or faces an awful threat. The insurmountable and imminent threat is sure to take over the world or destroy it unless the human race gets the better of it. The main characters display wisdom, genius, and heroics in their roles as a phenomenal scientist or technology expert. They eventually become heroes who save the world from the ordeal, which very few people had known about until then. Moreover, the menace that is targeting the world is generally made by humans themselves.
This is the kind of science fiction movie that we have been in for the past two years. The screenwriter is unknown, but we know the protagonist as “deepfakes,” a name that first appeared on Reddit.com. Many people may not be aware of the story; however, it is quite familiar to some. R&D groups of academics have been advancing deepfake synthetic-media-development methods for legitimate causes in health, sports, and fashion, and as a fun by-product of artificial intelligence (AI) technology.
Unfortunately, in the last couple of years, the whole thing has nearly spun out of control and paved the way for cyber-crimes. When the time comes, we will either encounter a cyber-attack and face the threat or protect the world. The only thing missing is the heroes to save the world. Those in the online world who are aware of the danger are hoping to meet the heroes in four months. The Deepfake Detection Challenge (DFDC) process will start in December and finish by the end of March 2020.
Cyber-security giants hunt deepfake at the DFDC competition.
DFDC is held by the world’s technology giants Facebook, Microsoft, and Amazon Web Services (AWS) Amazon’s cloud platformin collaboration with Partnership on AI, an AI consortium made up of 90 partners from 13 countries, including Apple, Google, Intel, IBM, and Sony. Evoking a Deepfake Safari, DFDC invites cyber-security technology developers around the world to come up with innovative technologies to detect deepfake or manipulated fake synthetic media. With an award of $10 million from Facebook and $1 million from AWS, the DFDC organization will determine the participants’ best cyber-security models in the deepfake synthetic media hunt by testing the participating cyber-security technologies on a giant synthetic media pool until the end of March. After completing the first data set on October 21 , DFDC is slated to start the nearly four-month-long competition with a launch at the Neural Information Processing Systems (NeurIPS) conference at the Vancouver Convention Center in Can ada on December 8.
R&D usually competes on the free market, but there is no time to wait.
Years-long R&D studies result in academic papers and literature in an endless race for innovation-focused concepts. During the commercialization stage, intellectual and industrial property rights are registered with a patent. Society gets to see the end product only when it has been put on the market as a new product or technology. The competition takes place only under market conditions. The ranking is decided by the consumers’ preferences and demand. The deepfake menace has turned into the greatest cyber-threat over just a couple of years, causing huge alarm and panic for the world as a game-changer in the R&D process. There is no time to wait for market conditions to mature for cyber-security.
Deepfake synthetic media steals the truth from people’s perception and judgment and makes them believe false truths by blinding their minds. It may reach a level to pose a threat to individuals’ rights, international and social peace, as well as the future of democracy in just a matter of months. Particularly the United States and China with billions of dollars of AI investments along with university technology departments and technology development companies in the EU, Russia, and Japan, are developing new deepfake technologies to help humanity. In doing so, they are greasing the wheels for intelligence, crime, and terror groups that could throw society’s perception of truth into chaos by using manipulated media as a weapon against people, as well as for computer hackers and malicious assailants who are bent on hurting a person or an organization. Many deepfake technologies are within easy reach and use for any purpose on mobile apps disguised as something else or on the interne t’s open-source platforms.
US presidential elections: a spectacular gala opportunity for deepfake.
Anyone familiar with the deepfake threat agrees that the US presidential elections in November 2020 provide a spectacular opportunity for the gala of a deepfake science fiction movie with a predictable screenplay and yet-to-be-decided cast. Also, by the end of 2019, software development platforms are set to start sharing the code blocks for manipulating the 60 human facial muscles on deepfake apps. And, by the end of 2020, these platforms will start sharing the results of the first studies and code blocks on Audio Cloning, which is set to solve the sound issues on deepfake media by adapting to English along with other languages.
The big guns of information technology want to make controlled progress before facing the first and biggest assault during the US presidential race, displaying determination through an urgent coalition with the DFDC organization to confront deepfake synthetic media before it has gained the strength to replace the truth. The competition is open on DFDC’s website (deepfakedetectionchallenge.ai) to contestants that deliver the deepfake cyber-security software codes they developed in December for testing, provided that they agree to the data set license, exclusively prepared and financed by the project partners for the event between December 2019 and March 2020. The website emphasizes that none of the content on the social media and video sharing platform will be on the data set. Deepfake media detection models created by the participants will be tested by the DFDC’s founding partners with a test set, and a mechanism will be created to score the results. Contestants will own the rights to the trained models on the training data set, and Facebook and project partners will acquire the rights to use these data sets from the participants as part of the competition.
A historic box office success is imminent for the first part of a science fiction movie with the whole online world watching. Let us hope for a happy ending. Enjoy watching!