Facebook’s new deepfake ban comes with a politician-sized loophole

We probably should have seen this coming. 

Yesterday, following months of criticism, Facebook announced a new “manipulated media” policy meant to curb the spread of deepfakes and misinformation on its gargantuan platform. And while the narrowness of the ban has already raised questions about its likely effectiveness, another perhaps more pressing problem has come to the fore: Politicians get a particular exception to the rule. 

That’s right, known media bad actors like Donald Trump will be able to continue to post edited and manipulated media as long as said media is not in the form of a paid post and Facebook deems it somehow newsworthy. Similarly manipulated content, under the new policy, would be removed by Facebook were it not posted by a politician. 

This caveat to Facebook’s bold proclamation was explained over email by company spokesperson Andrea Vallone.

“For non-ads – Politicians are held to the same standards as other users when it comes to manipulated media,” explained Vallone. “If a politician posts content that violates our manipulated media policy, we will remove the content. If the content is particularly newsworthy, we will evaluate whether it should be allowed on the platform despite the violation on a case-by-case basis.”

Catch that last sentence? We followed up to ask Vallone who determines whether manipulated media posts are newsworthy, and, if determined to be so and not removed, whether they will be labeled as such. We received no immediate response.

Importantly, Vallone corrected an earlier statement from Facebook communications manager Andy Stone. 

“On ads/promoted content, no,  content that violates our community standards in ads (of which this manipulated media policy is a part), whether posted by a politician or anyone else,” she wrote. 

It’s worth noting that the new policy, as it currently stands, only applies to an extremely specific category of content. Namely, in addition to video “edited or synthesized […] in ways that are not apparent to an average person, and would likely mislead an average person to believe that a subject of the video said words that they did not say,” the video also must be the “product of artificial intelligence or machine learning.”

Simply falling into one category or the other isn’t enough for Facebook to take action.

In other words, a deceptively edited video of Congresswoman Nancy Pelosi that makes her appear drunk is fine to remain on Facebook under the new policy. As is, likely, an edited video depicting Lindsey Graham “endorsing” the Green New Deal as machine learning did not appear to be involved in its creation. 

Oh, and remember, as we learned today, politicians are sometimes exempt from these new rules anyway. 

It’s going to be a fun election season.