Facebook says it is tightening its policies around misleading manipulated media by banning videos that are either heavily edited or are deep fakes
People share millions of photos and videos on Facebook every day, creating some of the most compelling and creative visuals on our platform.
About Facebook
Context & Ripple Effects
This ban is the endpoint of a slow policy crawl: Facebook spent 2017 swatting crude feed tricks like fake play buttons and static-image videos, then in 2018 extended fact-checking to photos and videos outright. By mid-2019 Zuckerberg was arguing that any deepfake definition was too broad and needed careful scoping — today's announcement is that scoping made into policy.
The catch, flagged in same-day coverage, is what the definition excludes: the distorted Pelosi video style of deception — slowing, splicing, selective editing — involves no synthetic media at all, so the new ban may not touch the most viral manipulation Facebook has actually faced.
First-order effects
Videos shown to be AI-generated or heavily edited beyond normal cuts now get removed under the misleading-manipulated-media policy rather than merely down-ranked or fact-checked.
Cheapfakes — slowed, cropped, or re-cut footage like the Pelosi clip — stay outside the removal rule and remain in the fact-checkers' lane, where labels slow distribution instead of stopping it.
Second-order effects
Creators of deceptive media have a clear incentive to shift from synthesis to editing, since the ban targets the expensive technique while leaving the cheap one largely intact.
Every rival platform inherits the same definitional fight: once Facebook publishes a line between 'heavily edited' and 'misleading,' advertisers and regulators will ask why other feeds draw theirs elsewhere.
Third-order effects
Enforcement hinges on detecting synthetic media technically, which pushes platforms toward provenance and labeling infrastructure rather than case-by-case moderation — the trust stack becoming policy plumbing.
If self-set definitions keep missing low-tech deception, the likely next actor is the regulator, turning a private scoping decision into a public standard for what counts as manipulated media.
The trend: Platforms are codifying synthetic-media bans faster than they can define manipulation itself, leaving the cheapest edits outside the rule while building detection infrastructure around the expensive ones.
Facebook is banning deepfakes, conforming a report from WaPo. This announcement comes ahead of plans for Facebook's Monika Bickert to testify at an @EnergyCommerce hearing about manipulated media on Wednesday. https://twitter.com/...
This is a fairly interesting shift from Facebook, which told me last year, as the altered Pelosi video rocketed around the Web, that “we don't have a policy that stipulates that the information you post on Facebook must be true” https://www.washingtonpost.com/ ...
.@facebook: We're going to ban meticulously altered video made to deceive the public. ALSO FACEBOOK: We will accept money for political ads with super obvious lies. https://twitter.com/...
Ahead of the 2020 US election, Facebook is now banning “all types of manipulated media”: content that “has been edited or synthesized in ways that aren't apparent to an average person and would likely mislead someone” https://about.fb.com/...
Pelosi spox @Drew_Hammill on Facebook's Deepfake ban: “Facebook wants you to think the problem is video-editing technology, but the real problem is Facebook's refusal to stop the spread of disinformation.” https://twitter.com/...
The most tech solutionist possible thing is banning deepfakes while explicitly changing your advertising policy to allow regular fakes https://twitter.com/...
Facebook consulted with many players on this, but the player emerging as the biggest target — the Biden campaign — isn't satisfied. The new policy offers only the “illusion of progress,” the campaign tells us. https://www.washingtonpost.com/ ...
Biden spokesman says Facebook's new manipulated-media rule offers only the “illusion of progress”: It “does not get to the core issue of how their platform is being used to spread disinformation” https://www.washingtonpost.com/ ...
By focusing on the (ultimately less important) “deepfakes,” Facebook affirms the (dubious) power and legitimacy of AI/ML systems of the sorts they also claim solve all problems—including the problems they create by not doing otherwise. https://twitter.com/...
Updated from last night: Biden campaign calls new Facebook deepfake policy “illusion of progress,” given the kinds of videos it would not target for removal. Click here for the full nuance of what is/isnt okay: https://www.washingtonpost.com/ ...
So Facebook is banning some deepfakes but there are big exceptions, like videos manipulated for the point of parody/satire and video edited to omit/change the order of words https://www.washingtonpost.com/ ...