Microsoft rolls out deepfake detector tool ahead of US election
By Tanisha Jindal
Microsoft, a tech giant, launched new software that can assist in spotting ‘deepfake’ content. The tool aims at spotting synthetic media (deepfakes) with the launch of a tool for evaluating videos and still photos to develop a manipulation score. This Authenticator software analyzes any photo or each frame of a video, scanning for evidence of manipulation that could not be visible to the naked eye.
Microsoft notified that it has cooperated with AI Foundation in San Francisco to form the "video authenticator" tool available to political campaigns, news outlets, etc. involved within the election process. Fake posts that seem to be real are of main concern in the upcoming US presidential election, particularly after fake social media posts erupted in a huge number during the 2016 election.
Microsoft explained that the Video Authenticator tool was formulated using a public dataset from Face Forensic++ and tested on the DeepFake Detection Challenge Dataset, which it remarks are both leading models for training and testing deepfake detection technologies. Microsoft has also affirmed that it created technology into its Azure cloud computing platform that allows developers of photos or videos to add data in the background that can be taken up to review whether imagery has been altered.
Microsoft is also operating with the University of Washington and others on assisting people to be savvier when it comes to spotting misinformation from reliable facts. According to Microsoft, Practical media knowledge can encourage us all to think critically about the context of media and become more active citizens while still admiring satire and parody.