One of the worst features of deepfake artificial intelligence systems is that they can produce a human voice even from a short recording. But a new software tool known as AntiFake can help prevent this. Among other things, you can use fake versions of people’s voices along with fake videos to make it sound like a politician or celebrity has said something they never actually said. spoke.

Additionally, there have been cases where people have received phone calls in the fake voice of a friend or family member asking them to send money as soon as possible due to an emergency. Additionally, duplicated voices can be used to bypass voice recognition security systems.

Although already to exist AntiFake, a technology designed to detect whether so-called legitimate voices are deepfakes, is reportedly one of the first systems to prevent the creation of such hoaxes. In short, it does this by making it much harder for AI systems to read basic vocal characteristics in voice recordings of real people.

“The tool uses a competing artificial intelligence technique that was originally part of the cybercriminals’ toolbox, but now we’re using it to protect against them,” said inventor Ning Zhang, an associate professor of computer science and engineering at the University of Washington in St. Petersburg. Peters. “We slightly distort the recorded audio signal, distort it, or distort it enough to make it sound accurate to human listeners, but that’s completely different from AI,” Louis said.

Even if it’s a deepfake was created The audio from a recording altered by AntiFake will not sound anything like the speaker’s actual voice. Testing so far has shown the technology to be more than 95% effective at preventing the synthesis of convincing deepfakes.

“While I don’t know what’s next in AI voice technology — new tools and features are constantly being developed — I believe our strategy of turning adversaries’ methods against them will continue to be effective,” Zhang said. Source