Facebook begins letting fact-checking partners review photos and videos and also lets them flag stories before being prompted
which ads, directed at which groups, were transmitted by FB in '16. http://newsroom.fb.com/... Alex Stamos / @alexstamos : Here is the transcript of our call this morning with reporters. http://twitter.com/... Kerry Flynn / @kerrymflynn : Facebook's Guy Rosen: “We do it because civic discourse is something we at Facebook strongly believe in, and we know it can thrive on our platform when it's safe.” Gavin Sheridan / @gavinsblog : Was there a lockdown? Facebook are now trying to aggressively mitigate. Something people have been calling on for years. Some good, some bad, some ok measures. The key is *never* believe what Facebook as a company says, only believe their actions. http://newsroom.fb.com/... Kerry Flynn / @kerrymflynn : TechCrunch @JoshConstine asks about fact-checking systems overseas and Tessa Lyons said something about they're working with Canada and Mexico on giving them more access to data / tools Kerry Flynn / @kerrymflynn : Tom advised that we should all stay on subject, as in DON'T MENTION CAMBRIDGE ANALYTICA. These are the experts in “election integrity” Thanks: @joshconstine See also Mediagazer
Context & Ripple Effects
This move extends a program that began with warning labels on stories flagged by two or more third-party checkers in late 2016, when Snopes and AP could only react to text links already surfaced to them. By March 2018 Facebook is widening both the format — photos and videos, where misleading content had largely escaped review — and the workflow, letting partners initiate flags rather than wait for user reports.
The timing matters: weeks earlier, the Mueller indictment showed how Groups tools were exploited by Russian actors, and Facebook's own detection work on coordinated amplification via fake accounts had established that manipulation ran through imagery and networks, not just shareable headlines. The company is under visible pressure from reporters and researchers to show its mitigation is keeping pace.
First-order effects
- Fact-checking partners such as Snopes and AP gain two new powers immediately: reviewing photo and video content, and proactively flagging suspect stories before user reports prompt them — shifting their role from reactive raters to active hunters.
Second-order effects
- Bad actors who relied on images and video to evade the 2016 text-labeling system now face review across formats, pushing coordinated campaigns toward channels with weaker checking — private Groups being the obvious adjacent surface exposed by the indictment.
Third-order effects
- Proactive flagging plus expanded formats turns fact-checking into standing platform infrastructure rather than an incident response — but later reporting that Facebook has overruled checker decisions on divisive content with little transparency shows the structural tension between delegated judgment and platform control that this expansion deepens.
The trend: Platform moderation is evolving from reactive, text-only labeling toward proactive, multi-format trust infrastructure run through third-party partners whose authority remains contingent on the platform's own discretion.