Community Notes, on X and rolling out on Meta's services, falls short of stopping misinformation; eliminating rewards for posting misinformation would help more
Musk's own posts have been successfully noted 167 times — but only 88 are currently still visible. — Almost a third of notes submitted in February were related to crypto scams or other ToS violations. Free labor for the world's richest man. … Adam Kucharski / @adamjkucharski : As I noted in my recent piece (www.bloomberg.com/news/article...), a key issue with any reactive fact-checking is speed. New data from below piece highlights this even more clearly: [embedded post] Dave Lee / @davelee.me : NEW and FREE to read at @opinion.bloomberg.com: — Today, Meta is starting to roll out its version of X's Community Notes for Instagram, Threads and Facebook. — Will it work? Analysis of 1.1 million examples of the crowdsourced fact-checking system show it's not stopping the spread of misinformation. Threads: Matthew Facciani / @matthewfacciani : Community fact checks are quick, and viewed as relatively trustworthy. However, but they often receive far less visibility than the false information they correct. Also, fans of public figures can coordinate efforts to have fact checks removed, undermining their effectiveness. https://www.bloomberg.com/... See also Mediagazer
Context & Ripple Effects
Community Notes was initially presented as a relatively transparent, crowd-consensus alternative to conventional fact-checking. But earlier reporting found that most contributed notes were never shown, while an election-claims review found only a small share of eligible corrections became public: most community-written notes remained invisible and many notes on false election claims did not clear the display threshold.
A recent large-scale study also showed that professional fact-checkers remain important source material within the ostensibly crowd-driven system. This analysis extends that record from note production to note effectiveness, just as Meta carries the model onto more services.
First-order effects
- Meta’s rollout to Facebook, Instagram, and Threads inherits a model in which corrections can arrive too late or fail to become visible, limiting their ability to change exposure to the original false post.
- On X, the findings sharpen the gap between submitting a note and delivering a public correction; coordinated opposition to notes can further determine which annotations remain visible.
Second-order effects
- Publishers, fact-checkers, and researchers may face greater demand to supply evidence into a system whose distribution rules—not merely its source quality—determine whether corrections reach users; professional fact-checkers already rank among frequently cited sources in notes.
- Scam and other policy-violating content may remain especially costly for platforms if reactive annotations substitute for removing the incentives or reach that make such posts worthwhile.
Third-order effects
- If platforms continue to replace or narrow proactive moderation with crowd annotation, trust-and-safety performance will increasingly hinge on ranking, visibility thresholds, and anti-coordination safeguards rather than the mere availability of fact checks.
- The broader policy question shifts from whether users can label misinformation to whether platforms redesign engagement and monetization systems that reward its creation and spread.
The trend: Crowd-sourced correction is becoming part of platform moderation infrastructure, but its value will depend on whether platforms pair it with incentive and distribution controls.