Deepfake Expert Warns There Will Be 'Perfect Authentic' Deepfake

This updated guide examines Deepfake Expert Warns There Will Be 'Perfect Authentic' Deepfake and organizes the essential facts, background, and practical takeaways in clear American English.

" It is still very easy to identify with most videos deepfake today," said Hao Li, associate professor of computer science at the University of Southern California. " But there are real examples, really Apparel ".

He said: " Soon, there will be a time when there is no way we can really detect deepfake anymore, so we will have to find other solutions."

Li created a deepfake video about Russian President Vladimir Putin at an MIT technology conference last week. According to him, this video is intended to demonstrate the current state of deepfake technology, which is growing faster than he thought. At that conference, he said that the " perfect and undetectable deepfake" videos " still" a few years away " were coming out.

But after CNBC asked more about that timeline in an email after the interview, Li said that recent developments, including the arrival of the hugely popular Chinese app, Zao, forced him. have to " recalculate" the timeline we give.

" In some ways, we already know how to make it," he said in an email. " It just needs to be trained with more data and complete."

Advances in artificial intelligence now allow deepfake videos to become more and more reliable, and it is also becoming more and more difficult to decode real videos from edited videos. That has raised alarm bells about the issue of misinformation, especially in the context of the United States and many other countries preparing to enter political elections in the near future.

Reference: BusinessInsider