BBC tests Intel ‘deepfake detector’
July 24, 2023
Intel believes it has a solution to deepfakes. Central to the system is a technique called Photoplethysmography (PPG), which detects changes in blood flow detected via the face -faces created by deepfakes don’t give out these signals, Intel told the BBC. The system also analyses eye movement to check for authenticity.
By examining both these traits, Intel believes it can work out the difference between a real video and a fake within seconds. The company claims FakeCatcher is 96 per cent accurate.
The BBC used a dozen or so clips of former US President Donald Trump and President Joe Biden for their test, some were real, some were deepfakes created by the Massachusetts Institute of Technology (MIT).
In terms of finding the deepfakes, the system appeared to be pretty good. The BBC mostly chose lip-synced fakes – real videos where the mouth and voice had been altered. It got every answer right, bar one.
But with the authentic videos problems were apparent: several times the system said a video was fake, when it was in fact real. The more pixelated a video, the harder it is to pick up blood flow. The system also does not analyse audio. So some videos that seemed fairly obviously real by listening to the voice were allocated as fake.
Intel told the BBC the system is being overly cautious. Better to catch all the fakes – and catch some real videos too – than miss fakes.
As with facial recognition systems, it seems that in lab conditions impressive results are achieved, but ‘in the wild’ the performance is more patchy.