Facebook says it took down 1.3B fake accounts from October-December and 12M pieces of COVID-19 misinformation; Facebook has 35,000 staff tackling misinformation
ReutersChavi Mehta
Context & Ripple Effects
Facebook's moderation reporting has become a quarterly ritual since its first transparency report in 2018 counted 583M fake accounts closed in a single quarter — a number that climbed to 2.19B by Q1 2019. The through-line is that removal volume keeps growing even as the company insists it is winning.
This report adds two new layers to that series: for the first time the headline numbers are dominated by COVID-19 content — 12M pieces removed — alongside 1.3B fake accounts from October to December, and Facebook names its enforcement headcount at 35,000 staff. The fake-account figure sits on top of the ML detection tool Facebook credited with taking down 6.6B accounts in the prior year.
First-order effects
Facebook's 35,000 moderators are now split between two enforcement fronts — identity fraud at billion-account scale and health misinformation — stretching the same review pipeline across qualitatively different problems.
The 12M COVID-19 removals mark health content becoming a standing moderation category for Facebook, no longer an episodic election-style push.
Second-order effects
Because most takedowns run on automated detection rather than human review, the marginal cost of another billion removals falls toward zero — but so does outside visibility into how many real users vs. bots those figures represent.
Advertisers and policymakers get a self-reported metric series with no independent audit, which raises the stakes for Facebook's later dispute over whether a dozen 'superspreader' accounts drove most vaccine misinformation — a claim the company rejected in August coverage of the 12 superspreaders episode.
Third-order effects
If each report shows larger takedown numbers without shrinking the problem, the durable answer platforms converge toward is verifying personhood itself rather than chasing accounts after creation — the space the proof-of-personhood concept addresses.
Self-published transparency reports risk hardening into an expected corporate disclosure format, inviting regulators to demand standardized, auditable moderation metrics across platforms.
The trend: Platform moderation is shifting from episodic cleanup campaigns toward continuous, machine-run enforcement at billion-account scale, with health misinformation emerging as a permanent category and verification of real humans as the endgame.
1/ Facebook announced today they deleted 1.3 billion fake accounts between Oct - Dec 2020. They're highlighting this as an example of them fighting disinformation effectively. Actually, it shows opposite and there's a big question about consumer fraud that needs to be asked.
@jesselehrich
Jesse Lehrich on x
an amazing line from @facebook: they belittle the notion that they have any financial interest in allowing misinfo... and then brag they're tackling it *at the expense of user growth & engagement.* https://about.fb.com/... https://twitter.com/...
@goangelo
Angelo Carusone on x
2/ RE Fraud - 1.3 billion is *a lot* of accounts. Think of all the advertisers on the platform (political orgs, civic groups, corporations). Facebook took money from advertisers, then FB showed those ads to fake accounts they let proliferate on the platform unchecked. It's fraud
@curtiskitchen
Curtis Kitchen on x
I feel more gross every day I think about how cool I thought their ad/campaign targeting capabilities were 5 years ago. Believed in them fully. https://twitter.com/...
@green_footballs
Charles Johnson on x
ONE POINT THREE BILLION fake accounts, with a capital B. Unreal. https://twitter.com/...
@dlauer
Dave Lauer on x
Did Facebook refund the money they obtained from advertisers by defrauding them? Of course not. Delete your facebook account. Stop supporting fraud, misinformation and extremism. Increasingly the answer to “where were they radicalized?” is “Facebook”. https://twitter.com/...
@slpng_giants
@slpng_giants on x
Nothing says, “we are doing something about disinformation” by deleting a bazillion fake accounts that were operating with their full knowledge through the last election season. https://twitter.com/...
@goangelo
Angelo Carusone on x
4/ Separately, this action doesn't show that FB was effective at addressing this problem; it's proof of how ineffective and negligent they have been. 1.3 billion accounts were able to proliferate...before they acted? They waited until October? Not earlier. C'mon.
@dog_congress
Dog Congress on x
@GoAngelo It also begs the question: if they found and deleted 1.3 billion fake accounts—how many more are there that they have not found and deleted? Surely they're not claiming they got rid of every last one?
@chiangku
Mike on x
if i were a paying advertiser, i'd be asking for one bigass refund right now https://twitter.com/...
@edbott
Ed Bott on x
He's right. Facebook is proudly displaying a metric that proves they are incapable of addressing disinformation effectively, and they're also systematically perpetrating fraud on their advertisers. https://twitter.com/...
@goangelo
Angelo Carusone on x
3/ RE Fraud (cont...) - Every FB advertiser should demand a refund for any money FB charged them to serve ads to these fake accounts. Only way Facebook will start addressing this on an ongoing basis and not once every few years is if it costs them money. A refund bare minimum.
@trekonomics
@trekonomics on x
I was told Big Tech was going to get the pandemic under control with apps and tracking people and stuff and instead we got facebook and insta spreading anti-mask and anti-vax conspiracies, literally killing people.
@bobcesca_go
Bob Cesca on x
And the Daily Wire Red Hats can post whatever the hell they want, including posts that violate FB's terms of service. As long as deals like this exist, FB will continue to be a corrupt data-mining shithole. https://twitter.com/...
@gregpinelo
Greg Pinelo on x
Put another, more accurate way: Facebook allowed the creation of over a billion fake accounts (that we know of) and belatedly deleted the ones it could identify without informing users who may have seen or shared fraudulent content. https://twitter.com/...
@rmac18
@rmac18 on x
Ahead of Zuckerberg's congressional testimony on misinformation this week, Facebook employees seem unclear on what is being labeled as COVID misinfo. Internally, one asked why FB applied a label typically seen on misinfo posts to this standard vaccine selfie from Michelle Obama. …