SOCHI, September 17. /TASS/. Protection from deepfake technology (fake videos with overlapping faces and voices on unrelated videos using Deep Learning neural networks) doesn’t exist yet, so given this serious threat, it would be dangerous to use biometric data, InfoWatch CEO Natalya Kaspersky said on Friday.
"You've probably heard about deepfake technology. Hacking experiments were conducted at Amazon and Microsoft hosting facilities equipped with biometric data. Amazon missed 75% of cases, while Microsoft let slip 66% of cases. Generally speaking, there is currently no defense system against deepfakes," she said.
Kaspersky explained that finding a face on the Internet and using it to make online payments is not difficult theoretically. " And that, of course, would be done," she said, clarifying that it would be difficult for someone to prove that they were not involved in such transactions.
She called for introducing criminal penalties for data leaks and suggested that people should not submit their biometric data until there is a reliable defense system against deepfakes.