Sep 26, 2018

Why academics are creating deepfakes

Illustration: Sarah Grillo/Axios

Deepfakes — realistic AI-generated audio, video, and images — are denounced as a threat to democracy and society by experts and lawmakers. So why are academics producing research that advances them?

Why it matters: Increasingly accessible tools for creating convincing fake videos are a "deadly virus," said Hany Farid, a digital-forensics expert at Dartmouth. "Worldwide, a lot of governments are worried about this phenomenon. I don't think this has been overblown."

Academic researchers are rapidly creating new methods for faking videos, photos, and audio. But they say their goal is not to destroy democracy, but to make new tools for creativity, and help improve other emerging technologies.

  • They call the technology "synthetic content generation."
  • In its benign form, researchers say, the techniques can be used in filmmaking, dubbing, or virtual reality, and also as training data to improve self-driving cars.
  • But they acknowledge that there is serious potential for harm when the technology is misapplied. In a paper published this summer, a pair of law scholars wrote:
"The volume and sophistication of publicly available academic research and commercial services will ensure the steady diffusion of deepfake capacity no matter efforts to safeguard it."
— University of Texas professor Bobby Chesney and University of Maryland professor Danielle Citron

Axios reached out to several academics who have published recent research that could be used to create deepfakes. Two responded.

  • Caroline Chan, an MIT graduate student who as a UC Berkeley undergrad created a system to simulate body movements in videos, said her research group has also worked on methods of detecting digital forgeries.
  • "As a community it is important to us to both advance the state of the art in content creation and be able to separate fake from real content with high confidence," she told Axios.

Aayush Bansal, a PhD candidate at Carnegie Mellon University, developed a technique to replace one person’s face with another's in a video.

  • But he said that it had positive as well as negative potential uses: One way of improving systems that aim to detect faked videos is by pursuing new ways of generating them, he said.
  • "Since these new approaches essentially work by learning a model of what real data looks like, they are also very good at detecting fake content that was manipulated in any way or created from thin air," said Chan.

Go deeper:

Go deeper

Tech can't remember what to do in a down market

Illustration: Rebecca Zisser/Axios

Wall Street's two-day-old coronavirus crash is a wakeup alarm for Silicon Valley.

The big picture: Tech has been booming for so long the industry barely remembers what a down market feels like — and most companies are ill-prepared for one.

Brace yourself for a coronavirus outbreak

Illustration: Aïda Amer/Axios

Public-health officials’ warnings about the coronavirus are sounding increasingly urgent, with one top CDC official asking the public yesterday "to prepare for the expectation that this might be bad."

Reality check: Other administration officials, including President Trump himself, were more subdued in their assessments. But underneath those tonal differences, the reality of the coronavirus is the same: It spreads quickly, and has already spread to many countries, making it likely to start spreading here, too.

Go deeperArrow2 hours ago - Health

Exclusive: Pro-Trump group plans post-Super Tuesday blitz on Democrats

Democratic presidential hopefuls take the debate stage in South Carolina. Photo: Logan Cyrus/AFP via Getty Images

Pro-Trump super PAC America First Action is preparing to unleash a series of targeted, swing-state attacks on the Democrat most likely to face President Trump after Super Tuesday, people familiar with the group's plans tell me in an exclusive preview of its strategy.

The state of play: The group has been tracking favorable/unfavorable ratings in Florida, Georgia, Michigan, North Carolina, Ohio and Pennsylvania for 2020 candidates Joe Biden, Bernie Sanders, Elizabeth Warren, Pete Buttigieg and Michael Bloomberg — under the theory that if Trump wins each of these six states he would win re-election.