In an impressive effort to differentiate the artificial from the real, Microsoft has launched a deep-fake detector tool called the “Video Authenticator” to add to their growing list of innovative technologies. The objective of the tool is to spot synthetic or artificially created media and with the help of algorithms analyses and generates a certain ‘manipulation score’.
Further information was disclosed by Microsoft with them saying, “In the case of a video, it can provide this percentage in real-time on each frame as the video plays. It works by detecting the blending boundary of the deep-fake and subtle fading or grey-scale elements that might not be detectable by the human eye.”
This technology is however not just limited to videos, but also content pieces and news articles the subject of testing which has been co-developed by the San Francisco based AI Foundation. The need to detect malicious intent and a surge to purposely disinform people rather than inadvertently misinforming, is now more than ever. Therefore, to not trick the audience, the detection technology is at work with critical thinking and intelligence to determine and identify such deep-fakes.
Interestingly, this is going to be rolled out as a massive exercise ahead of the 2020 Presidential Elections in the United States. At a time where ‘fake news’ and ‘misnomers’ in the campaign propaganda will be rife with inaccuracies, the Microsoft detection technology can help authenticate and run trials to identify which source or journalist’s content is being supplied with false information – possibly with the intention of taking down a rival candidate.
If this tech turns out to be a success, then it would yet again be a feather in the hat for Microsoft and so, artificial manipulation may have a fierce counterpart in order to let true and genuine videos and content lead the charge.
Do check out: