Interview with nine members of the Facebook team tasked with fighting misinformation on how recent product changes affect the way the News Feed works
Nicholas Thompson / Wired : See also Mediagazer
Context & Ripple Effects
Facebook's misinformation apparatus has been built in public stages: first [[a:914963|warning labels and lower News Feed ranking for stories flagged by partners like Snopes and AP]] in late 2016, then detection of coordinated amplification using fake accounts ahead of the US election fallout, then a May 2018 package of academic research proposals, a short film, and a news literacy program.
This Wired interview with nine members of the misinformation team is the most granular look yet at how those stages translate into actual News Feed mechanics — and it lands just weeks before the News Feed head would publicly defend keeping InfoWars on the platform, making the gap between the team's engineering and the company's line-drawing the live question.
First-order effects
- The nine-person team's product changes directly re-rank distribution inside the News Feed, so false stories that once needed a human flag now face algorithmic demotion before any label appears.
- Third-party fact-checkers like Snopes and AP shift from being the trigger for visible warning labels to feeding signals into ranking decisions made at scale by this team.
Second-order effects
- Publishers and page operators now have to optimize against a feed whose suppression rules are set by an internal misinformation team rather than published policy, raising the value of the earlier relevancy-score transparency the News Feed team described to Slate in 2016.
- Academics responding to Facebook's request for proposals gain a formal channel into measuring these changes, turning independent research into a de facto audit layer over the team's work.
Third-order effects
- The pattern points toward platforms governing speech through ranking rather than removal — content stays shareable but loses reach — with the InfoWars defense showing the ceiling of that approach when the content itself, not its spread, becomes the controversy.
- If the structure holds, misinformation policy effectively gets written by small internal product teams plus contracted fact-checkers, prefiguring today's fights over who audits platform algorithms and under what access.
The trend: Platform misinformation control is migrating from post-hoc human labeling toward ranking decisions embedded in the feed itself by dedicated internal teams, with external researchers and fact-checkers recast as signal providers rather than gatekeepers.