Facebook has recently taken a position of not removing faked videos, choosing instead to reduce their reach and display them alongside fact-checking information. Now a series of viral deepfake videos of Facebook CEO Mark Zuckerberg and other famous individuals is putting Facebook’s policies to the test.
Motherboard reports that artists Bill Posters and Daniel Howe recently partnered with the advertising Canny to create AI-generated fake videos of people talking about a mysterious “Spectre,” which is actually an exhibition the artists put on at the Sheffield Doc Fest in the UK.
“Imagine this for a second: One man, with total control of billions of people’s stolen data, all their secrets, their lives, their futures,” the faked Zuckerberg says. “I owe it all to Spectre. Spectre showed me that whoever controls the data, controls the future.”
The video doesn’t exactly put Zuckerberg or Facebook in a good light, but the Facebook-owned Instagram is declining to scrub it from its service.
“We will treat this content the same way we treat all misinformation on Instagram,” an Instagram spokesperson tells Motherboard. “If third-party fact-checkers mark it as false, we will filter it from Instagram’s recommendation surfaces like Explore and hashtag pages.”
Canny tells Motherboard it created the videos using a CannyAI software’s Video Dialogue Replacement (VDR) technology by training for 12 to 24 hours on 20- to 45-second scenes of each person’s face.
Researchers at Samsung recently demonstrated how realistic talking heads can be created by AI using a single still portrait of a person, and this technology will only get better and more realistic over time.