Ad company tests Facebook policies with deepfake video of Zuckerberg posted to Instagram

Ad company tests Facebook policies with deepfake video of Zuckerberg posted to Instagram
© Greg Nash

An ad company posted a deep fake video of Facebook founder Mark ZuckerbergMark Elliot ZuckerbergWarren warns Facebook may help reelect Trump 'and profit off of it' Martin Luther King Jr.'s daughter knocks Zuckerberg for invoking her father while defending Facebook Hillicon Valley: Zuckerberg defends handling of misinformation in political ads | Biden camp hits Zuckerberg over remarks | Dem bill would jail tech execs for lying about privacy | Consumer safety agency accidentally disclosed personal data MORE to Instagram last week to test the social media giant's moderation policy.

The video was created by artists Bill Posters and Daniel Howe in partnership with advertising company Canny, according to Vice News.

Zuckerberg is seen giving a sinister speech about Facebook's power but with audio and video that is fully synthetic.

"Imagine this for a second: One man, with total control of billions of people's stolen data, all their secrets, their lives, their futures," Zuckerberg appears to say in the deepfake video. "I owe it all to Spectre. Spectre showed me that whoever controls the data, controls the future."eo is from a September 2017 address Zuckerberg gave about Russian election interference on Facebook.

The original video is from a September 2017 address Zuckerberg gave about Russian election interference on Facebook.

One of Canny founders, Omer Ben-Ami, told Vice that to create the deepfake video,and many others they have posted, the ad company used a proprietary AI algorithm, trained on 20 to 45 second scenes of the target face for between 12-24 hours.

For the Zuckerberg video, engineers clipped a 21-second segment out of the original seven minute video, trained the algorithm on this clip as well as videos of a voice actor speaking and then reconstructed the frames in Zuckerberg's video to match the facial movements of the voice actor.

"The true potential we see for this tech lies in the ability of creating a photo realistic model of a human being," Ben-Ami told Vice.

"For us it is the next step in our digital evolution where eventually each one of us could have a digital copy, a Universal Everlasting human. This will change the way we share and tell stories, remember our loved ones and create content."

The video of Zuckerberg comes shortly after a fake video of House Speaker Nancy PelosiNancy PelosiDemocratic senator rips Trump's 'let them fight' remarks: 'Enough is enough' Trump touts Turkey cease-fire: 'Sometimes you have to let them fight' Mattis responds to Trump criticism: 'I guess I'm the Meryl Streep of generals' MORE (D-Calif.) was widely shared on Facebook.

The video, viewed over 2 million times, did not use any deepfake technology and instead just slowed the California lawmaker down to make her appear unwell or potentially drunk.

Facebook announced it would not delete the video, and instead de-prioritize it in people's newsfeeds.

Neil Potts, Facebook’s director of public policy, said at the time that if a manipulated video of Zuckerberg like the one of Pelosi was posted it would stay up.

“We will treat this content the same way we treat all misinformation on Instagram. If third-party fact-checkers mark it as false, we will filter it from Instagram’s recommendation surfaces like Explore and hashtag pages," an Instagram spokesperson said.

The video of Zuckerberg included a fake chyron from CBS News to make it appear like it was a broadcast.

The news station has "requested that Facebook take down this fake, unauthorized use of the CBSN trademark,” a CBS spokesperson told The Hill.

Updated Wednesday at 9:11 a.m.