Facebook Cracks Down on Deepfakes - Sort Of

By Gavin Whenman on at

Facebook has announced it is banning misleadingly manipulated videos - i.e. deepfakes - although, because it's Facebook, it's not banning all of them and people are pissed off at the exceptions.

From this week, the social network will remove media from its platform if:

  • It has been edited or synthesised – beyond adjustments for clarity or quality – in ways 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
  • 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.

However, Facebook has insisted it won't remove deepfakes that are supposed to be parody or satire, nor will it ban videos have "been edited solely to omit or change the order of words." Those are some pretty wide exemptions and they mean, for example, that the doctored video of Labour's Keir Starmer that the Conservatives produced during the general election wouldn't be removed, although the ability to share it on the platform might still be restricted under Facebook's existing anti-fake news procedures.

As tech reporter Chris Stokel-Walker has pointed out, the new policy's exemptions also mean malevolent deepfakers can still publish whatever they want to Facebook and claim it is satire to sidestep the ban. This (admittedly not convincing) deepfake video of Facebook founder Mark Zuckerberg, for example, wouldn't be affected as the artists behind it claim it is a parody.

Other social media platforms are also considering how they will handle the scourge of misleadingly manipulated videos, with Twitter soliciting feedback in November last year on its draft anti-deepfake policy, although it has yet to unveil the final rules. [Facebook via The Next Web]

Featured image: Facebook