Agoracom Blog

Synthetic media: The real trouble with #deepfakes – SPONSOR: Datametrex AI Limited $

Posted by AGORACOM-JC at 5:22 PM on Monday, March 16th, 2020

SPONSOR: Datametrex AI Limited (TSX-V: DM) A revenue generating small cap A.I. company that NATO and Canadian Defence are using to fight fake news & social media threats. The company announced three $1M contacts in Q3-2019. Click here for more info.

Synthetic media: The real trouble with deepfakes

By M. Mitchell Waldrop

  • The snapshots above look like people you’d know. Your daughter’s best friend from college, maybe? That guy from human resources at work? The emergency-room doctor who took care of your sprained ankle? One of the kids from down the street?
  • “Deepfakes play to our weaknesses,” explains Jennifer Kavanagh, a political scientist at the RAND Corporation and coauthor of “Truth Decay,”

Nope. All of these images are “deepfakes” — the nickname for computer-generated, photorealistic media created via cutting-edge artificial intelligence technology. They are just one example of what this fast-evolving method can do. (You could create synthetic images yourself at Hobbyists, for example, have used the same AI techniques to populate YouTube with a host of startlingly lifelike video spoofs — the kind that show real people such as Barack Obama or Vladimir Putin doing or saying goofy things they never did or said, or that revise famous movie scenes to give actors like Amy Adams or Sharon Stone the face of Nicolas Cage. All the hobbyists need is a PC with a high-end graphics chip, and maybe 48 hours of processing time.

It’s good fun, not to mention jaw-droppingly impressive. And coming down the line are some equally remarkable applications that could make quick work out of once-painstaking tasks: filling in gaps and scratches in damaged images or video; turning satellite photos into maps; creating realistic streetscape videos to train autonomous vehicles; giving a natural-sounding voice to those who have lost their own; turning Hollywood actors into their older or younger selves; and much more.

Deepfake artificial-intelligence methods can map the face of, say, actor Nicolas Cage onto anyone else — in this case, actor Amy Adams in the film Man of Steel.

Yet this technology has an obvious — and potentially enormous — dark side. Witness the many denunciations of deepfakes as a menace, Facebook’s decision in January to ban (some) deepfakes outright and Twitter’s announcement a month later that it would follow suit.

“Deepfakes play to our weaknesses,” explains Jennifer Kavanagh, a political scientist at the RAND Corporation and coauthor of “Truth Decay,” a 2018 RAND report about the diminishing role of facts and data in public discourse. When we see a doctored video that looks utterly real, she says, “it’s really hard for our brains to disentangle whether that’s true or false.” And the internet being what it is, there are any number of online scammers, partisan zealots, state-sponsored hackers and other bad actors eager to take advantage of that fact.

“The threat here is not, ‘Oh, we have fake content!’” says Hany Farid, a computer scientist at the University of California, Berkeley, and author of an overview of image forensics in the 2019 Annual Review of Vision Science. Media manipulation has been around forever. “The threat is the democratization of Hollywood-style technology that can create really compelling fake content.” It’s photorealism that requires no skill or effort, he says, coupled with a social-media ecosystem that can spread that content around the world with a mouse click.


Tags: , , , , ,

Comments are closed.