UTSA computer scientists work to combat photoshopped videos called Deep Fakes

SAN ANTONIO – A deep fake is basically a photoshopped video. It’s a video that was fabricated. An example could be where someone else’s face and voice was edited over another person, and these videos can be very convincing and may appear real. Deep fakes are already out in the world, and so now schools, professors, and computer scientists are already working to combat them.

“It’s increasingly important and it’s not going to go away,” Dr. Amanda Fernandez at the UTSA department of computer science said.

Fernandez teaches deep learning and artificial intelligence, and part of her subject matter is deep fakes.

“Right now, it is relatively easy to generate a deep fake that is believable and that very few people will have a way to identifying whether this is real or not,” student Michael Geyer said.

Geyer is working in the computer science department to create programs to identify these counterfeit videos. There’s the cliche “you can’t believe everything you read”, but now you can’t believe everything you see or hear.

“Anybody with a laptop and maybe a few hours of their time they’re willing to dedicate it could create anything they would like to use for the election coming up. So it is extremely important to be vigilant about the source of your news,” Fernandez said.

The scary part is Fernandez said these videos could become more and more popular, and more prevalent on social media with the 2020 election around the corner.

“We will do the primary research to create these models, really detect and defend them and they’ll become the cybersecurity tools of tomorrow,” Fernandez said. “The end goal of this work would be to create not only a platform that any person can use on their own, but also something integrated into these platforms that scans through these videos and automatically alerts you by saying, ‘hey, this video was deep faked.’”

About the Author: