Facebook will not remove deepfake video of Mark Zuckerberg from Instagram

12 Jun 2019

Facebook profile of CEO Mark Zuckerberg. Image: grinvalds/Depositphotos

In what has proven to be a spectacular challenge to its own espoused community standards, Facebook has said that it will not remove a deepfake video of its own co-founder from Instagram.

When a doctored video of US House speaker Nancy Pelosi began making the rounds on social media, Facebook declined to remove it. The company’s head of global policy management, Monika Bickert, said at the time that it did not violate the company’s community standards despite being demonstrably false (it was slowed down to make Pelosi appear drunk in what was a rather crude and unsophisticated fake).

At the International Grand Committee on Big Data, Privacy and Democracy in late May, Estonia’s Keit Pentus Rosimannus grilled Neil Potts, a Facebook global policy director, about what would happen if, say, hypothetically, a doctored video of Mark Zuckerberg appeared on the platform. Would the company leave it there? To which Potts definitively said: “Yes.”

Now, that thought experiment has been catapulted in reality as a deepfake video of Facebook co-founder Mark Zuckerberg that was uploaded to Instagram (which Facebook owns) will not be removed.

“We will treat this content the same way we treat all misinformation on Instagram,” a spokesperson for the company confirmed. “If third-party fact-checkers mark it as false, we will filter it from Instagram’s recommendation surfaces like Explore and hashtag pages.” This is similar to how the video for Pelosi was handled by the firm.

The company has also said that it currently uses image detection technology to find content that has been debunked by its third-party fact-checking program.

Vice first reported that two artists from an advertising company, Bill Posters and Daniel Howe, had created the fake in partnership with advertising company Canny. The video portrays Zuckerberg sitting at a desk waxing lyrical about the powers that he has inherited grace á a suitably ominous-sounding fictional organisation called Spectre, likely a nod to the James Bond film.

“Imagine this: one man in control of billions of people’s stolen data. All their secrets, their lives, their futures. I owe it all to Spectre. Spectre showed me that whoever controls the data controls the future.”

The audio is of an actor doing a rather robotic impression of Zuckerberg. The video is a modified version of a real statement the Facebook CEO gave regarding Russian election interference on the Facebook platform, doctored using Canny AI’s ‘video dialogue replacement’ technology.

This deepfake is part of an exhibition that ran at the UK’s Sheffield DocFest, which also featured doctored videos of Kim Kardashian and Donald Trump, all discussing their involvement with Spectre.

Facebook was previously excoriated by commentators for apparent hypocrisy when it removed Zuckerberg’s messages from recipients’ inboxes despite the fact that this function is not available to the average user. When questioned about the move, the company claimed it was done in the name of corporate security, though some argued that failure to inform the recipients of this action amounted to a breach of user trust.

So, perhaps the decision to treat the Zuckerberg fake with the same approach as the doctored Pelosi video is a bid to ensure it cannot be accused of having two different sets of rules for its own executives and the general public.

This revelation will, as one may expect, likely further amplify the debate around how deepfakes could be leveraged to manipulate the 2020 US presidential election, similar to how data harvesting and manipulation on Facebook by Cambridge Analytica were alleged to have undermined the 2016 election.

Facebook profile of CEO Mark Zuckerberg. Image: grinvalds/Depositphotos

Updated, 11.02am, 12 June 2019: This article was updated to include comments from an Instagram spokesperson.

Eva Short was a journalist at Silicon Republic