We all know Game of Thrones and its effect on popular media. One of the characters in the series, the assassin who belonged to the cult of “Faceless Men”. That character parallels the emergence of “Deepfakes” and his ability to change his face is nothing short of deception. Deception is the main notion behind deepfakes. Deepfakes have been on the rise recently. Video manipulation and deception have been possible for a long time. The emergent technology has made it very easy to automatically capture a person’s image and put it onto someone else.
What Are Deepfakes?
Deepfake is an AI-powered method for human image synthesis. Through this technique, fake videos and images can be created. Existing images and videos are combined and superimposed onto source images or videos. This is done using a machine learning technique known as a generative adversarial network. Deepfake videos are hard to differentiate by untrained eyes because they can be truly realistic. They are generated for various reasons that can be personal revenge, to manipulate financial markets or to disrupt international relations.
Deepfakes and FakeApp
In January of 2018, a proprietary desktop app was launched called FakeApp. Through this app, the users can easily create, edit and share videos with faces swapped. The deepfake software employs the AI Framework TensorFlow of Google. Although celebrities are the main focus of such fake videos, it becomes imperative to question the authenticity of every video or image online.
Other software and apps have superseded the FakeApp. It is becoming increasingly easy to replace the face of one person in a video with another person’s face. Due to a number of scandals including celebrities and deepfakes of political figures like Trump, it becomes necessary to create automated ways to detect these Deepfake videos.
The Role of Face Verification
Face recognition is an advanced biometric technology that is capable of identifying and verifying a person through their facial profiles. The software works by comparing and analyzing the unique facial features of people. It is mostly used for security purposes but there are quite vast applications of the tool.
How does Face Recognition Work?
Before proceeding further, you should know how facial verification solutions work. The algorithms of biometric face recognition follow a number of stages which include:
- Capture: The first step includes the collection of a sample of facial records during a certain period.
- Extraction: Then the gathered data is used to make templates based on them.
- Comparison: The collected data is then cross-referenced with the existing templates.
- Matching: Then the algorithm determines whether the facial features of a new sample match with the one from the facial database or not.
How Are Deepfakes Threatening Facial Verification?
Although it has been long believed that “seeing is believing,” this is not the case anymore. Videos portraying people doing and saying things they never did or said are causing people to disbelieve everything they see online.
Although deepfakes are harming the content you see online, it is also causing immense damage to verification systems put in place. Facial verification solutions are used as security measures for a number of online accounts, banks, and even smartphones.
Through deepfakes, it has become very easy to trick the face recognition software. These days, financial institutions and banks employ facial recognition technology for security purposes. Through online face verification, users make their accounts secure. As bio metrics of every human being is unique, through the employment of face verification solutions, user accounts can be made more secure. Deepfakes are a threat to this security measure. With deep fakes, fraudsters can hack into anyone’s account by tricking the facial recognition software.
The currently employed two-dimensional facial recognition method includes a person holding some form of identification like a driver’s license up to their while staring at the camera. This is even more problematic because if this information gets hacked, there are two forms of Personally Identifiable Information (PII) that can be compromised. This makes identity theft way easier to pull off.
How to Prevent This?
The creation and widespread usage of deepfakes and deepfake technology highlight a number of issues including privacy concerns, the authenticity of content online and the rise in online hacks and identity thefts. Detecting deepfakes is considerably a difficult task as the fraudsters and fakers will continue to get better at creating deepfakes. The research into the detection of deepfakes has to always keep up and even get a bit ahead.
Recently Social Media giant Facebook has launched its own deepfake detection project. The main purpose of the project is to train the Artificial Intelligence model, which will better assist in the detection of deepfakes. Other than this, a lot of research is being carried out and experts predict that they will soon be able to make systems that will be able to identify deep fakes.
Technical Content writing is my passion. I have broad experience in writing for the technical field. I started my writing at the age of sixteen when I was in college. Now I’m already writing for contribution sites as an independent influencer. I wrote many articles on medium and many of them are published in various publications. I wrote many case studies for businesses and nowadays working on Facial Recognition technology to let them know the real need for digital transformation for their business.