Meta Brings Community Notes to LatAm Without Showing What Success Looks Like

When Meta announced this month that it was beginning to test Community Notes in sixteen Spanish-speaking countries in Latin America, it brought some numbers. After a year and a half in the United States, the company said Community Notes has more than 1.4 million contributors, who have produced more than 50,000 notes across Facebook, Instagram and Threads. In July alone, more than 200 million people in the US saw a note.

Those figures are meant to signal success. But in fact, they highlight some problems inherent in community notes and they raise more questions than answers.

The contributors are arriving. The notes are not.

The most difficult to interpret figure by itself is the 1.4 million. It tells us that, a year and a half in, the contributor base has grown. But the ratio of published notes per contributor is out of whack: 0.036. This ratio is likely higher for notes written per contributor, but due to the bridging mechanism that requires bipartisan agreement, of those written notes only a few get published. From data about X’s community notes, we know that the share of published notes hovers around 10%.

There are innocent explanations. Any volunteer system has a participation curve with a long tail. But this ratio also raises questions of who actually writes notes. We know from X that only a small number of contributors write most notes, a power distribution as we see with many other phenomena on the internet. If that was also the case for Meta, one of their key metrics had to be taken with a grain of salt.

The ratio also raises a question about the long-term viability of the approach: If contributors don't see their notes published, they might lose interest in writing more. Indicator has documented a similar development for X. They reported that just eight AI contributors wrote 50.3% of all visible Community Notes on the platform in the first three weeks of May 2026.

The number that matters. And a baseline to compare it to.

The most relevant figure is the 50,000 notes published. Is that a lot? To make sense of it, we need a baseline for comparison. Unfortunately, I'm not aware of any data on the scale of the former fact-checking program in the US over a similar timeframe.

We do have data from the EU. According to Meta's own reports submitted under the Code of Conduct on Disinformation, the aggregated number of content that was assigned a fact-checking label in the EU is 49,340,000 between 1 January 2025 and 31 December 2025. Granted, that's a bigger market, but the timeframe is shorter than the one covered by the Community Notes numbers. So it's 50,000 against almost 50 million.

Why such a discrepancy? The most likely explanation is the design of the programs. Community notes are assigned on a per-post basis: even reincarnations of the same post — duplicates, copies — have to go through the entire annotation process. Someone has to write a note, and it has to be rated helpful by (usually) disagreeing users before it's displayed. Fact-checking labels, by contrast, are also assigned to matching claims: once automated systems find duplicates of a debunked claim, they can apply the label without any delay.

This is an important mechanism to counter the slop that is increasingly flooding users’ feeds. EFCSN members have repeatedly highlighted the problem where hundreds of pages automatically post fabricated or misleading stories about politicians, celebrities, or social issues.

What does it mean to view a note?

The most eye-catching figure in the announcement is the 200 million.

More than 200 million US users viewed a community note in July alone, Meta says. That is a remarkable share of the country. It signals scale.

Meta might say that this is evidence that 200 million people were informed. To me, it shows that 200 million people encountered content misleading enough to attract a note that survived a cross-partisan consensus test. That context arrived in a smaller font, partially hidden behind a “more” link and only by clicking "sources" and then a link in a pop-up can the reader get the full picture.

Furthermore, considering that notes take a long time to appear, the misleading content was viewed even more often, as usually most of a viral post's impressions accrue in its first hours.

This is compounded by a policy choice: Meta has stated that content receiving a Community Note does not receive the distribution penalty that accompanied many fact-check labels. Under the fact-checking program that notes replaced in the US, a ‘false’ rating reduced a post's reach. A note does not. That means Community Notes function primarily as an information layer rather than an enforcement mechanism. Meta has not presented them as a system that automatically penalizes repeat offenders.

Meta also did not provide any information on which notes did not get seen. Experience and research on X show that the consensus system effectively gives partisans a veto over notes they don’t like. Completely accurate notes pointing out clear falsehoods on very viral posts are routinely suppressed because partisans can’t reach consensus, even with thousands voting. A system that lets partisans censor information they don’t like might be a great way to avoid offending those partisans but it is not an effective way to tackle false or misleading information.

Notes written by US Community Notes users that do get approved also don’t show up for Facebook, Threads and Instagram users in the rest of the world, effectively allowing false information from US publishers to go unchallenged in the rest of the world since Meta discourages fact-checkers in other countries from labeling posts by publishers outside their own country.

Testing a model without defining success.

Meta is now testing Community Notes in Argentina, Bolivia, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay and Venezuela. Users can sign up as contributors; no notes will be publicly visible during the test; independent fact-checking continues in the meantime.

On its face, this is a cautious, staged approach. The obvious question is what criteria Meta will use to decide whether the test has succeeded. And nowhere in this announcement does Meta state what result would cause it to keep professional fact-checking in any of these sixteen countries.

What publication rate would be adequate? What time-to-visible-note would be fast enough to matter in an election week? What level of coverage of viral false claims would count as sufficient? What degree of contributor concentration, or evidence of coordinated rating, would trigger Meta’s Trust and Safety teams to step in?

The announcement addresses none of these questions. It says, “We will closely monitor the results and evaluate performance country by country, taking local needs into account. Full rollout in a given country, including starting to publish Community Notes and gradually migrating the independent verification program, will occur when we are confident that the product is working correctly.”

Taking into account how little data Meta has shared thus far, the decision whether to cancel professional fact-checking and roll out Community Notes might as well be completely arbitrary. There is nothing the public can hold Meta to account for.

A substitute that doesn’t substitute.

Meta's own Oversight Board has pointed out an important aspect in this regard, writing, “Any decision to phase out fact-checking, as referenced in Meta’s recent announcement, should not be based on the existence or success of community notes since they are designed for different outcomes, and community notes often rely on fact-checking. We urge Meta not to consider community notes a direct replacement for fact-checking. "

The Indicator’s Alexios Mantzarlis backed this point up with a preliminary analysis. He found that just 3% of Instagram posts fact-checked by AFP’s US bureau since April 2025 have a corrective note. This is an indication that combining professional fact-checking with a community-based system makes a lot of sense.

Show numbers that matter!

But actually none of this is really about metrics, data, or evidence. It is about transparency and accountability. Meta is rolling out a program with policy choices that change how misinformation is addressed; the Oversight Board has warned that Community Notes should not be treated as a direct replacement for fact-checking; and much of the data needed to independently assess the program's performance remains with Meta

Holding power to account requires the ability to see what power is doing. A handful of companies shape what billions of people encounter as information, and transparency obligations in Europe have made it possible to scrutinize some aspects of their operations.

If Community Notes are going to become the new standard for adding context on Meta's platforms in Latin America, the public should be able to see what "working correctly" actually means.

Disclosure: The European Fact-Checking Standards Network (EFCN) has received funding from Meta on one occasion in the past.