Logo
Logo
Search
UPGRADE
ABOUT
RESOURCES
REPORTING
LOGIN

A first look at Meta’s Community Notes

Zuckerberg's pivot to crowdsourcing features three jacked dudes, a flirty farmer, and some useful notes

Alexios Mantzarlis
Alexios Mantzarlis

Jun 17, 2025

A first look at Meta’s Community Notes

In January, Mark Zuckerberg announced a major shift in Meta’s approach to content moderation that included retiring its fact-checking program and launching a Community Notes feature inspired by X.

This was a craven political decision, but Community Notes is not without benefits. Crowdsourced labeling of misleading content can in principle serve as a multiplying force for professional fact-checkers, while also introducing additional legitimacy and transparency to platform moderation.

Meta launched its Community Notes pilot in the US on March 13, phasing out its fact-checking program in parallel.

Three months later, the company has yet to share any updates about the pilot. Despite promising to “be transparent about how different viewpoints inform the Notes displayed in our apps,” Meta has shared nothing like the database of notes and ratings that X published at the very beginning of its program (then called called Birdwatch).

The company claims it is committed to sharing more data within an unspecified timeframe.

To get a sense of what’s going on, I found three Facebook and Instagram users who are in the Community Notes pilot and asked them to screenshot what they saw over the course of four days. I collected a total of 61 individual posts and annotated them. Paid subscribers can access my spreadsheet with all the details at the bottom of this post.

Based on the screenshots shared by my panel of users, it’s clear that Community Notes on Facebook and Instagram is not yet ready for prime time. 

While several notes helpfully flagged content that was misleading, only a handful provided truly valuable context. Some were inaccurate and potentially AI-generated. Meta’s reliance on user flagging on posts that may benefit from a note appears particularly shaky at this stage. Posts in my sample had minimal reach and in some cases raised concerns about privacy and proportionality.

This is obviously not a representative analysis of the program, but absent any kind of data from the company, it’s also all we can do. Let’s dig in.

logo

Upgrade to read the rest.

Become a paying member of Indicator to access all of our content and our monthly members-only workshop. Support independent media while building your skills.

Upgrade

A membership gets you:

  • Everything we publish, plus archival content, including the Academic Library and our Resources
  • Live monthly workshops and access to all recordings

Keep Reading



Indicator is your essential guide to understanding and investigating digital deception.

cursor-click