WASHINGTON (FOX 5 DC) - As fake videos designed to spread misinformation become easier to create, some tech companies are working to ensure what you’re seeing is the real deal.
Videos known as “deepfakes” are essentially high-tech face swapping and can make it appear people are doing and saying things they never did.
Mounir Ibrahim, who works for a San Diego startup called Truepic, says as photos and videos become easier to doctor, image verification will eventually become standard throughout society.
Truepic is a system that verifies a picture or video is authentic and unaltered from the instant the capture button is pressed. It’s technology that’s caught the attention of lawmakers in D.C.
“We’re meeting with a variety of officials on the Hill,” Ibrahim told FOX 5 after a meeting at the State Department. “Members of the House and members of the Senate and their staffers, but also with the State Department and a variety of other offices in the U.S. government.”
Part of why Truepic’s technology is of interest to the government has to do with growing concerns about the impact deepfake videos could have on elections and the difficulty lawmakers have had trying to legislate the issue. There are currently no laws on the books specific to deepfakes.
While there are ways to detect fake videos, Ibrahim compares that to a cat and mouse game.
“As good actors identify fraudulent videos, bad actors will constantly innovate to work around the latest detection mechanisms,” he said.
Instead, Truepic creates a new kind of image: one that you can say for sure is authentic. It’s comparable to the verification check you see on Twitter and Facebook that prove someone is who they claim to be.
A tech company out of the UK called Serelay has created a similar system and CEO Roy Azoulay showed us how it works.
“This is a real photo I took in London a few months ago,” Azoulay explained. “It was a peaceful little demonstration.”
He showed us how the system reacted to a doctored version of the picture with a knife photoshopped in it.
“Within 30 seconds, we can say, alright there is a problem here,” Azoulay said. “This is not an exact match and these are the areas (of the photo) we think there could be a problem with.”
This kind of technology is already being used by insurance companies to verify that accident photos and damage is legitimate.
As fake videos become more common so will new ways to verify what’s real.