(CN) – On the eve of a congressional hearing on media manipulation and deception, Facebook said it will remove digitally manipulated videos – commonly called “deepfakes” – from its platform.
Monika Bickert, Facebook’s head of global policy management, announced the new policy in a blog post Monday, saying Facebook will remove videos that have been misleadingly edited to depict people saying things they didn’t say. It will also remove media with content that has been superimposed or merged onto a video to make it look authentic.
The policy will not apply to satire or parody videos, or to video manipulated with AI tools to remove or change the order of words. A team of fact-checkers will still review said media.
“If a photo or video is rated false or partly false by a fact-checker, we significantly reduce its distribution in News Feed and reject it if it’s being run as an ad. And critically, people who see it, try to share it or have already shared it will see warnings alerting them that it’s false,” Bickert wrote.
“This approach is critical to our strategy and one we heard specifically from our conversations with experts. If we simply removed all manipulated videos flagged by fact-checkers as false, the videos would still be available elsewhere on the internet or social media ecosystem. By leaving them up and labeling them as false, we’re providing people with important information and context.”
Facebook confirmed to the news service Reuters today that it will not take down the Pelosi video since it does not meet the new standards for manipulated media.
The announcement comes as Bickert is set to testify Wednesday before the House Subcommittee on Consumer Protection and Commerce at a hearing titled “Manipulation and Deception in the Digital Age.”
First Amendment advocates have raised concerns about censoring online content. Last year, a coalition of news broadcasters and civil rights groups signed on to a letter opposing California’s new law banning deepfakes, saying it unfairly attacks political speech and news reporting.
One of those groups, Electronic Frontier Foundation, also expressed concerns about congressional efforts to draft deepfake-regulating legislation as potentially overbroad.
David Greene, EFF’s civil liberties director, said in an interview Tuesday that he’s interested in whether Facebook will be able to apply its policy accurately.
“Content moderation is impossible to do perfectly and really hard to do well. Our concern is always that in the past we’ve seen lots of things taken down that shouldn’t be taken down and things that should be taken down to stay up there – and the decisions seem to be inconsistent and not transparent,” Greene said.
“What remains to be seen and what we need to be watchful for is whether they’ll be able to implement this in the narrow way it’s written,” he continued. “Is it going to catch the right stuff or is it going to catch the wrong stuff?”
Angst over political deepfakes and their unpolished cousin “shallowfakes” blew up last year with the online proliferation of two doctored videos – one that made House Speaker Nancy Pelosi sound drunk, and another cut to make presidential candidate Sen. Joe Biden sound like he made comments later criticized as racist.
“Only videos generated by artificial intelligence to depict people saying fictional things will be taken down. Edited or clipped videos will continue to be subject to our fact-checking program. In the case of the Pelosi video, once it was rated false, we reduced its distribution,” Facebook said in a statement.
In an email, Facebook spokesman Andy Stone said earlier reports that it would not ban deepfakes from political ads are inaccurate. “Whether posted by a politician or anyone else, deepfakes are not permitted in ads on Facebook,” Stone said.
Along with the new policy, Facebook said it will partner with Reuters to offer journalists free online training in spotting deepfakes.