Ready or Not, Mass Video Deepfakes Are Coming

Comment

It was mainly for fun that Chris Ume created a fake Tom Cruise.

The special effects artist wanted to try something different during the doldrums of 2020, so working with a Tom Cruise lookalike, he used AI and facial mapping technology to invent a series of comedic deepfake videos and , in early 2021, unleashed them on TikTok. The DeepTomCruise account quickly rose to popularity, then disappeared from the public mind, replaced by the next viral hijack.

Ume is now back and on a mission to commercialize video deepfakes for the planned metaverse and make them as essential to digital life as tweets and memes.

He will take the next step on Tuesday when a deepfake developed by Metaphysic, the company he formed with entrepreneur Tom Graham, makes it to the semi-finals of NBC’s “America’s Got Talent” reality show.

“It’s a good opportunity to raise awareness and show what we can do,” Ume said.

“We believe the web would be much better off if, instead of avatars, we lived in the hyper-real world,” Graham added, describing users’ ability to manipulate real faces with Metaphysic.

The startup’s appearance in front of millions on TV will lay the groundwork for its new website which aims to make it easier for everyday people to say and do things in their face that they never have. do in real life. Many other such sites are aimed at programmers and researchers.

And the act – in which they will follow a raucous appearance in the preliminary round which had them superimpose the face of a young Simon Cowell on the screen above a stage performer so that the judge appeared to be singing to himself – will offer an advertisement brilliant for a technology that is becoming more democratic with astonishing rapidity.

Still, some critics are horrified by this celebratory moment on a top TV show. Video deepfakes, they say, blur the line between fiction and reality that is barely clear now. If disinformation peddlers can be so successful with doctored words and images, imagine what they can do with full video.

“We are rapidly entering a world where everything, even videos, can be manipulated by just about anyone,” said Hany Farid, a professor at the University of California at Berkeley and an expert on deepfakes. “What can go wrong? »

The unveiling of what for most weeks this summer is the most-watched show on network TV comes at the end of a frenzied summer in the world of deepfakes, which use deep intelligence learning artificial to create fake media” or “AI-generated”).

While many Americans happily indulged in quaint analog activities like going to the beach, a start-up named Midjourney offered an “AI art generation,” in which anyone with a basic graphics card could, with a few keystrokes, create amazingly real images. To spend even a few minutes with – there’s Gordon Ramsay burning in his Hell’s Kitchen; Here’s Gandalf’s shredding on a guitar – it’s uncovering technology that makes Photoshop look like Wite-Out. Midjourney has amassed over a million users on its Discord channel.

Three weeks ago, a start-up called Stable AI launched a program called Stable Diffusion. The AI ​​Image Generator is an open-source program that, unlike some competitors, places few limits on the images people can create, leading critics to say it can be used for scams, misinformation policy and privacy violations.

“We should be worried. I follow technology everyday, and I am worried,” said Subbarao Kambhampati, a professor at Arizona State University’s School of Computing & AI who has studied deepfakes and virtual identities. He said he expects the “AGT” moment to see even more platforms like these take off, even as the technology improves day by day.

“It’s going so fast. Soon anyone can [to] create a moon landing that looks like the real thing,” he said.

Ume and Graham say deception is not their goal. Ume emphasizes entertainment value: the company will market itself to Hollywood studios that want to feature deceased actors in movies (with permission from an estate) or have artists play against their younger selves.

As for ordinary users, Ume says Metaphysic’s goal is to make online interactions more real – without the fancyness of video games or the dullness of Zoom. “I imagine being able to have breakfast with my grandparents in Belgium from here in Bangkok and feel like I’m really there,” Ume said from her current base.

Graham adds that synthetic media, far from harming privacy, will enhance it. “I would like to see a world where online communication is a more human experience owned and controlled by humans,” said Graham, a Harvard-educated lawyer who founded a digital graphics company before turning to crypto and , possibly, deepfakes. “I don’t think that’s happening in today’s Web2 world.”

Farid is not convinced. “They only tell half the story – the one about you using your own image,” he said. “The other side is someone else using it to defraud, spread misinformation and disrupt society. And you have to wonder if being able to move around a little more on Zoom is worth it.

Deepfake technology began eight years ago with the use of “generative adversarial networks”. Created by computer scientist Ian Goodfellow, it essentially pits two AIs against each other to compete for the most realistic images. The results were far superior to basic machine learning techniques. Goodfellow went on to work for Google, Apple, and now DeepMind, a Google subsidiary.

At first, deepfakes were used by skilled exploiters, who infamously grafted the faces of actresses onto porn videos. But with the technology requiring fewer tools, it can now be deployed by ordinary people for a range of uses, which Metaphysic hopes to explore further.

Earlier this year, the company attracted a $7.5 million investment from the Winklevoss twins, the social media-turned-crypto entrepreneurs, and Section 32, the venture capital fund of the original Google Ventures founder. , Bill Maris. “We believe the impact will be significant,” Section 32 managing partner Andy Harrison said of Metaphysic. Harrison, also a Google veteran, said he sees video deepfakes not as a threat but as an invigorating shift in consumption and communication.

“Honestly, I’m pretty excited,” he said. “I think it’s a new era in entertainment and social interaction.”

Critics, however, worry about the “liar’s dividend,” in which a web awash in video deepfakes muddies the waters even for legitimate videos, resulting in no one believing anything.

“Video was the final frontier of online verification. And now it could also disappear,” Farid said. He cited the unifying power of the George Floyd video in 2020 as unlikely in a world awash with deepfake videos.

Asked about “AGT’s” role in promoting deepfakes, a spokesperson for production company Fremantle declined to comment for the story. But a person close to the show who requested anonymity because she was legally prohibited from commenting on an ongoing contest said she believed there was a social purpose to the metaphysical act. “By using the innovation in a totally transparent way,” the person said, “they’re showing a mainstream audience how this technology can work.”

A solution to the truth problem could take the form of authentication. A cross-industry effort involving Adobe, Microsoft, and Intel would verify and make transparent the creator of each video to assure people it was real. But we don’t know how many would adopt it.

Kambhampati, the ASU researcher, said he fears the world will end up in one of two places: “Either no one trusts what they’re looking at anymore, or we need an elaborate authentication system for them to do so.”

“I hope it’s the second,” he said, then added, “it’s not that great either.”

Comments are closed.