Deepfakes are one of the most concerning developments when it comes to AI. They are often horrifyingly realistic videos of people saying and doing things they’ve never done before. Generally, they are made by superimposing a face, usually one of a celebrity, over that of another person to create these deceptive videos.
They are often meant to be funny, or at least not made with serious intentions. Some want to make movies with famous actors (opens in new tab) without ever having to leave the house. But as you can imagine, there is a ton of porn that uses deepfakes (opens in new tab)which range from fantasies created with video game characters to problematic depictions of real people.
Aside from the blatant disrespect this can often show to the images of people used, one of the most dangerous things about deepfakes is that they are really good. I’ve seen deepfakes I never thought weren’t real, and that’s especially dangerous when we have an internet that can spread disinformation like wildfire.
“Deepfake videos are everywhere now. You’ve probably already seen them; videos of celebrities doing or saying things they’ve never actually done. explains Ilke Demir, senior staff researcher at Intel Labs.
It turns out it takes an AI to catch an AI, and thankfully Intel has been working on just that. The company has been working on efforts as part of its Responsible AI work, developing what Intel has dubbed the FakeCatcher. This is technology specifically designed to detect fake videos like deepfakes, and it can reportedly do so in milliseconds with an accuracy rate of 96%.
Intel explains that many other tools that try to detect deepfakes try to analyze the raw data in the files, while FakeCatcher is trained with deep learning techniques to detect what a real human looks like in videos. This means it’s trained to look for all the little things that make people real. Subtle things like noticing blood flow in pixels of a video, which the deepfakes apparently haven’t mastered yet. Although this may be how they learn to control our human blood effects. That’s a creepy sentence if I’ve ever written one.
FakeCatcher will operate on a web-based platform that will hopefully allow anyone to access the technology. On the back end, it uses Intel’s 3rd Gen Xeon Scalable processors in conjunction with a bunch of proprietary software to do the required calculations.
If you want to learn more about Intel’s FakeCatcher, watch the video at the top of the article. Demir will also host a Twitter Spaces event on November 16 at 11:30am PST where she will elaborate on the technology used. That is if Twitter’s new owner, Elon Musk, doesn’t fire the people (opens in new tab) that ensure that Twitter Spaces works for its time.