Goodbye DeepFake: Intel develops AI to detect fakes on networks

Intel is accelerating the resolution of issues related to processors and graphics cards, and is committed to rapid information management through the introduction of artificial intelligence to quickly recognize data. Now, as part of its social responsibility, the semiconductor brand is introducing FakeCatcher, a technology capable of detecting 96% of deepfake videos.

Intel said in a statement that it is important to provide the world with a deep counterfeit detection tool: “Video deepfakes are everywhere these days. You have probably already seen them; Videos of celebrities doing or saying things they’ve never actually done.” says Ilke Demir, senior researcher at the Intel lab.

With face detection algorithms optimized with multi-threaded software libraries applied to real-time video and image processing tools, Intel was able to analyze “blood streams,” the elements that make us human, versus “computer eyes.”

The function of FakeCatcher, one of the first of its kind, is to detect these subtleties in human faces and process them into spatio-temporal maps to use deep learning to determine whether a video is fake or not.

DeepFakes: A Growing Threat in the World

Intel’s report includes data from Gartner that warns of $188 billion in total spending on cybersecurity solutions that prevent deepfake tools from maliciously infiltrating the digital environment. Simply put, the discovery process is currently costly and time consuming.

Enabling FakeCatcher will help social media platforms block potentially harmful videos, just as global news outlets will be able to detect fake videos and thereby prevent unverified news from being published, the company says.