A computer scientist is working to develop a tool/app to identify the fake videos or footage of celebrities and politicians.
In an era full of fake videos and news stories, you might except some video proof to have a clear image of truth.
A Google engineer Supasorn Suwajanakorn has almost developed such an app, and demonstrated his self created software while providing it 14 hours of training for footage of US’s former president Barack Obama during TED conference held in Vancouver.
The software uses a neural learning network to automatically mimic the way a person talks by observing that how human brain works, interpret the subject’s teeth, analyse the mouth movement while observing facial expression. By analyzing these things provides a perfect lip-sync.
Supasorn Suwajanakorn said “These results seemed intriguing, but at the same time troubling; it concerns me, the potential for misuse,”. Further he added “We don’t want it to be in the wrong hands, so we have to be very careful about it.”
Scientist is also spending his efforts with AI foundation on an app which is called as ‘Reality Defender’, that would be able to run automatically in web browsers in order to spot and flag fake pictures, videos or stories.
While addressing AFP he said “Video manipulation will be used in malicious ways unless counter-measures are in place,”. Moreover he explained “We have to make it very risky and cost-ineffective.”
This technology can build a moving 3D model of anyone’s face—and make it look like they’re saying any sentence—just by analyzing a large collection of photos. What could go wrong? Supasorn Suwajanakorn #TED2018 pic.twitter.com/mTfMBIjGbq
— TED Talks (@TEDTalks) April 11, 2018
No doubt Suwajanakorn’s software is remarkably behaving in tremendous way but still there is a ‘long way to go’ before it can perfectly measure the people, he also admitted.