McAfee LLC, an American global computer security software firm headquartered in Santa Clara, California, has reportedly launched ‘McAfee Deepfakes Lab’ to provide social media and traditional organizations with advanced AI (Artificial Intelligence) analysis of alleged deepfake videos. These videos are proliferating on the internet and can be used to spread misinformation about organizations and individuals during the 2020 U.S President Election season and beyond.
Sources cite that the new AI-based solution will leverage McAfee’s data science expertise and tools that integrate deep learning and computer vision techniques to uncover hidden patterns and recognize manipulated video elements that play a pivotal role in authenticating original media files. Social media platforms and news organizations can submit suspected videos to the Deepfakes Lab and obtain detection scores and heatmaps that explain the company’s evaluation of the videos’ authenticity.
Speaking on which, Steve Grobman, SVP and CTO at McAfee said that the 2020 U.S President Election season brings with it several critical concerns around election security and misinformation. Cutting-edge AI technologies are making it possible for broad fields of potential bad actors to develop deepfake videos that portray candidates as saying mislead information, he states.
Such deepfake videos could be launched during the election season and influence voters in a particular direction, Mr. Grobman said while adding that the company’s new solution leverages its AI capabilities to help media outlets recognize and counter misinformation before it can manipulate voters.
Deepfake videos are developed by the superimposition of existing videos, audio, and images, onto source media files by leveraging a cutting-edge deep learning technique named GAN (Generative Adversarial Networks (GANs). These GANs are used to synthesize fake videos that are indistinguishable from original ones. This technology can be used to develop politically motivated video content that could possibly spread across social media platforms before concerned individuals and trustworthy news organizations could challenge the video’s claims.