Two days before Facebook is set to appear in front of lawmakers at a House Energy and Commerce hearing on manipulated media, the social network has announced it’s banning all forms of deepfakes. The announcement represents a significant step forward for Facebook, which has been struggling to mend its ailing image with the 2020 presidential election right around the corner.
In a blog post, Monika Bickert, Facebook’s vice president of global policy management, said the company will take down videos that have been “edited or synthesized in ways that aren’t apparent to an average person” or are the “product of artificial intelligence or machine learning that merges, replaces or superimposes content onto a video, making it appear to be authentic.”
However, the policy doesn’t apply to parodies, satire, or clips “that has been edited solely to omit or change the order of words.” This means it also doesn’t apply to awidely viewed doctored video of House Speaker Nancy Pelosi, as it’s not forged by A.I. and instead, is merely the result of readily available speech software.
In addition to keeping an eye on A.I.-generated content, Bickert says Facebook is also “partnering with academia, government, and industry to expose people behind these efforts.”
As A.I. tech continues to advance, deepfakes — most of which are nearly impossible to identify — have emerged as one of the most critical challenges for governments and platforms. With a looming presidential election, companies have scrambled to come up with ways to crack down on and spot deepfake videos at scale. In November, Twitter also announced a new draft for deepfake policies, although it has yet to come into effect.
Unlike its peers, Facebook has been in the public crosshairs as it refuses to ban and fact-check political ads.
Last month, a number of organizations and companies including Facebook, Microsoft, the Massachusetts Institute of Technology, University of Oxford, and more joined hands to launch the Deepfake Detection Challenge, which aims to catalyze the development in deepfake detection through large data sets, awards, and research grants.
- Inside the rapidly escalating war between deepfakes and deepfake detectors
- Facebook’s new image-recognition A.I. is trained on 1 billion Instagram photos
- Summon this amazing Twitter bot to colorize your black-and-white photos
- Chess. Jeopardy. Go. Why do we use games as a benchmark for A.I.?
- Why teaching robots to play hide-and-seek could be the key to next-gen A.I.