X Implements New Measures to Detect Manipulation of Community Notes

X quietly rolled out an important update to its Community Notes program last week, with the implementation of a new process that can detect coordination between Community Notes contributors.

As explained by X:

“We’ve extended Community Notes ability to detect coordinating contributors with additional features targeting coordination between note writers and raters. When the scoring algo detects anomalous correlations in ratings, it automatically prevents those ratings from contributing to notes achieving helpful status.”

In other words, when the Community Notes system detects coordination between contributors to upvote or downvote a note, it will now treat the coordinated ratings as if they all came from the same user.

So, theoretically at least, that should dilute the capability for people to manipulate the Community Notes system, by coordinating with each other to mass support, or reject, notes for ulterior purpose.

Which is a key concern. Various studies have found that X’s Community Notes process is indeed being manipulated, with organized groups of contributors collaborating to get rid of notes that they don’t want X users to see.

Which is one of the reasons why 85% of all Community Notes are never displayed to X users, with notes contributors failing to reach consensus on the need for a note, or simply working together to hold back a note where such doesn’t align with their perspective.

And given that Community Notes are only displayed when contributors of opposing political viewpoints agree that one is necessary, many, many notes which should be shown, and which do provide valuable context, don’t get displayed.

That means that a lot of the most divisive political topics are free to proliferate on X, because on certain issues, people from opposite ends of the political spectrum will never agree.

For example, notes on issues like voter fraud, gender debates, the U.S. government’s tariffs, abortion, all of these issues are highly divided, and won’t ever reach full agreement on the need for clarification, based on ideological perspectives.

Which means they go unchecked, for anyone to read on X. And because they’re not Community Noted, users likely think that these are true, because that’s the way that X has framed it, that X is the best source of truth because the people vote on what’s acceptable, and what’s not.

But that’s not the full picture, while again, mass manipulation of the notes system also harms its overall credibility.

But X’s team is working on fixes, and has regularly implemented new approaches to improve the Community Notes system.

Hopefully those are having a valuable impact, and if they are, hopefully Meta will also look to implement the same changes in its Community Notes approach.