Deepfake technology is a powerful AI tool, one that’s been used — to varying degrees of success — for wrongdoing. In the face of the technology’s spread online, Intel FakeCatcher, a real-time deepfake detector, has been released.
Deepfakes are AI-powered faceswaps that use trained models to replace a person’s face in a picture or video. Previously, the faceswaps had to be done over long periods of time on existing footage, but trained models can now be used in real-time.
Experts have warned that real-time versions of this technology will be used in dangerous ways. For example, online catfishing on video chat services is now possible with the tech as well as a litany of other crimes.
With this thought in mind, Intel FakeCatcher sims to detect deepfakes online. Intel claims that the software has a 96% success rate, but how does it actually work?
FakeCatcher is said to work by analysing the “blood flow” of pixels. Created by Intel Labs senior staff researcher Ilke Demir, the software measures the amount of light that is absorbed and reflected by a person’s skin.
The software analyses the blood flow across 32 points of a person’s face to create a PPG map. This map is then analysed in real-time to determine whether or not the person you’re talking to is wearing a virtual mask.
Intel FakeCatcher is part of a massive initiative at the tech company to remove disinformation from the internet. With AI, disinformation is becoming an even bigger problem than normal, and Intel is aiming to fight against it.
“FakeCatcher is a part of a bigger research team at Intel called Trusted Media, which is working on manipulated content detection — deepfakes — responsible generation and media provenance,” Demir told VentureBeat. “In the shorter term, detection is actually the solution to deepfakes — and we are developing many different detectors based on different authenticity clues, like gaze detection.”
Detecting deepfakes has not been an easy task. Although only a small number of crimes have been committed with the AI tech, they are growing. In fact, it’s believed that Deepfake Scams will become a major issue as the technology continues to evolve.
It’s hoped that detection software like FakeCatcher will help to mitigate the upcoming wave of deepfake crimes. After all, if software like this helps to stop one person from being scammed or abused, it’s worth it.