Home > Business > Microsoft’s new deepfake detection tool rates bogus videos with a confidence score

Microsoft’s new deepfake detection tool rates bogus videos with a confidence score

95 Views

Deepfakes are now widely used by a range of people across the world including news agencies, politicians, and even meme-makers. So, it’s equally important for organizations to step up and build tech that can detect these AI-powered bogus videos manipulated to look realistic

Just ahead of the US Presidential elections, Microsoft has introduced new tools to detect manipulated videos that can help political campaigners and media organizations.. The detection tool, called Microsoft Video Authenticator, will analyze the videos frame-by-frame to give you a confidence score to indicate if the video was modified. 

The algorithm has been developed with the help of Microsoft’s Responsible AI team and the Microsoft AI, Ethics, and Effects in Engineering and Research (AETHER) Committee. And it’s designed to identify manipulated elements that are not easily caught by the human eye. The team trained the AI by using publicly available datasets such as Face Forensic++, and was tested on the DeepFake Detection Challenge Dataset.