Facebook is taking new steps to crack down on so-called “deepfake” videos, but won’t ban some forms of the videos entirely from its social-media platform.
Facebook said that it will now remove videos that appear to have been changed to make it look as if a person is saying something they actually didn’t say. Such videos, known as deepfakes, have been a source of controversy for Facebook, as many have involved politicians, including House Speaker Nancy Pelosi, D-San Francisco, who was the subject of deepfake video last year that was doctored to make it look like she was slurring her speech at an event in Washington, D.C.
But, politicians aren’t the only public figures to be the target of deepfakes. Last summer, a doctored video of Facebook Chief Executive Mark Zuckerberg appeared on Instagram in which he appeared to give credit to his success to a group called “Spectre” made the rounds and raised interest in the matter of deepfakes.
“There are people who engage in media manipulation in order to mislead,” said Monika Bickert, Facebook’s vice president of global policy management, in a company blog post announcing the policy changes. “While these videos are still rare on the internet, they present a significant challenge for our industry and society as their use increases.”
Bickert said a deepfake video will be subject to removal if it has been edited beyond reasons for clarity or quality that “aren’t apparent to an average person and would likely mislead someone into thinking that a subject of the video said words that they did not actually say,” and if it is “the product of artificial intelligence or machine learning that merges, replaces or superimposes content onto a video, making it appear to be authentic.”
Jeff Pollard, vice president and principal analyst at Forrester Research, said that while Facebook may have good intentions regarding its position on deepfake videos, he believes the company is engaging in what he calls “policy theater.”
“Itss true that Facebook needed to do something here, given that the Facebook is used to amplify and distribute content,” Pollard said. “The policy itself leaves plenty of loopholes from a content moderation perspective, and there’s no reason to think Facebook will get things right this time given its history.”
Not all deepfake videos will fall under Facebook’s new guidelines. Bickert said videos used for satire or parody, or where words have been cut out or had their order changed, will still be allowed on its sites.
Rory Lynch, a reputation and privacy attorney with London-based law firm Vardags, said Facebook’s new policies are likely to put pressure on other social-media outlets to tighten up on deepfake videos on their sites.
“The modus operandi of deepfakes is deception,” Lynch said. “The political and public relations implications of this increasingly sophisticated technology is also a serious risk factor for platforms. I expect to see others follow suit and track into Facebook’s slipstream on this issue.”
