How Community Notes Reduce Viral Misinformation | Keith Coleman, Jay Baxter | TED

0:04 – 2:41

A random internet user corrects the White House

The conversation opens on a striking image: a post about Iran claims the USS Lincoln has been damaged and shows supposed evidence, but the image is AI-generated. On the right sits the key intervention, a box that says, “readers added context they thought people might want to know.” Jay Baxter explains that this note is not written by a staff moderator or a famous expert. It is written by a regular contributor, and it only appears after people from different perspectives rate it as helpful. That detail matters because Community Notes is not just trying to label falsehoods. It often adds context to things that are technically correct but still misleading. Keith Coleman pushes that principle further: every post is eligible, including posts from heads of state, from X itself, and even from Elon Musk. He gives the most memorable example in the talk: in at least one case, the White House removed a post and issued an updated statement because of a note written by, as he puts it, “a random person on the internet.” Audrey Tang sharpens the image by adding that she heard a teenager caused one such retraction. Coleman calls that “a superpower for people.” That is the unexpected angle of the whole system. Most moderation tools feel top-down, like a referee blowing a whistle. Community Notes is closer to giving the crowd a shared annotation pen, but only when the crowd proves it can agree across disagreement. That is why people trust it more than a generic misinformation warning. The note itself is usually specific. It points to what is wrong, cites sources, and explains the issue in plain language. Instead of saying, “be skeptical,” it says, in effect, “here is the exact seam where this claim comes apart.” Tang frames it as a “context-engine for news,” and that phrase captures the deeper ambition. The product is not just trying to catch lies. It is trying to make public conversation less fragile by giving ordinary people a way to attach shared reality directly to viral posts.