AI deepfakes have the ability to hijack any living person’s biometrics and are an “existential threat to liberal democracy”, an expert in the field has chillingly warned.
Nina Schick, author of Deep Fakes and the Infocalypse, told Sky News today (September 14) of the imminent threat the new technology poses.
Deepfakes enable people to place one image of a person onto a video of someone else speaking.
The technology has faced widespread criticism for its role in celebrity pornographic videos, internet hoaxes and financial fraud.
But Nina has revealed one of the most alarming things is that it “isn’t just someone in the public eye that can be targeted”.
“If there is an image of you, or a piece of audio, it can be taken and your biometrics can be hijacked by AI,” she told Sky News’ Adam Boulton.
“I can train the AI to look like you and to sound like you and you don’t need much training to do it.
“Inevitably it’s going to get much easier so everyone is a target.”
Nina, who is also a political broadcaster, said the “technology is already here” and the only way to fight it is for a “society-wide inoculation”.
“There is going to be some technical tools we can deploy,” she continued.
“This is essentially a human problem. It is about building society-wide resilience in an age where everything you see can be faked and everything can be denied.
“Which is, obviously, an existential threat to liberal democracy.”
It comes as Daily Star Online revealed how deepfakes have soared during the coronavirus pandemic with criminals using the ever-popular Zoom app to create fake videos.
Source: Read Full Article