Meta opened Community Notes testing in 16 Latin American countries as a planned replacement for third-party fact-checking

Filed

September 29, 2026, 06:00 UTC

Actor

Meta Platforms, Inc., operating Community Notes on Facebook, Instagram and Threads

Target

Spanish-language users of Facebook, Instagram and Threads in sixteen Latin American countries, including several with documented coordinated-disinformation or high-conflict information environments

Source

Identified Openly sourced: the speaker is who they say they are

Confidence

documented Primary documents, court records, official disclosures

Techniques

Manufactured consensusBothsidesismDisclosure theaterAstroturfing

Where

Argentina, Mexico, Venezuela, Nicaragua, El Salvador, Ecuador

Channels

• Facebook
• Instagram
• Threads
• Meta Community Notes contributor pipeline
• phase-out of third-party fact-checking labels
• Meta Transparency Center product page

Summary

Analysis

Meta is not hiding the change. Its own Community Notes page, updated 9 September 2026, states that contributors in sixteen named Latin American countries may now sign up, that published notes are not yet shown in those markets, and that a full launch will include transitioning away from the third-party fact-checking program. That sentence is the load-bearing method. A professional check is a named institution making a claim under its own name, chosen through the platform's own vetting process. A community note is a crowd product that becomes visible only after enough raters who previously disagreed still endorse it. Each model rests on a selection process the audience does not see.

The Oversight Board's 10 September co-chair statement, which AFP moved into wider circulation on 12 September, does not accuse Meta of running a secret campaign. It says crowd notes and fact-checking serve distinct purposes, that notes often rely on fact-checks, and that Meta should not treat one as a replacement for the other. The Board restated earlier 2026 guidance: repressive regimes, polarized political violence, large-scale disinformation networks, high-risk elections, and active crisis or conflict are poor first markets. Those conditions are not theoretical on this list. The Oversight Board's risk factors, it says, apply to several markets on this list. El Salvador and Ecuador fit the risk profile the Board's guidance just described. Meta is opening a contributor funnel that is still invisible in those markets while announcing a later end to labeled checks; the Oversight Board lists those markets as poor first choices.

Why this is propaganda research rather than a product review: the doctrine in this cabinet is about methods that bypass or overwhelm evaluation. A fact-check label is a claim with a named owner. A community note appears only when a bridging algorithm finds raters who previously disagreed endorse it, so its owner is the ranking rule and its raters. The Board and IFCN argue that swap weakens accountability. Critics warn of bothsidesism if a documented record and a rumor share the same empty slot until raters fill it. Meta's own stated case for the swap, made when it ended third-party fact-checking in the United States in January 2025, was that the program had drifted into one-sided calls on contested topics and had cost the company more user trust than it built; Meta has since applied that same rationale to its international rollout. The Oversight Board's and IFCN's cautions recorded in this file answer that case; they do not erase it. The About page still mentions fact-checking partners while the product roadmap says those partners go away; a reader should check the dated roadmap. Astroturf capture is a risk the Board's criteria point to; no capture is documented in these markets. Coordinated networks already buy reach; they can also buy or assign note writers and raters. The method does not require Meta to intend a factional win. It only requires that correction travel more slowly than the post, and that the correction look ownerless.

Scoring the method, not the firm: the same swap on a state television app, a party forum, or a rival platform would be the same swap. Last two briefs this date covered commercial influence shops. This cycle rotates to a platform. Meta's openness about the test is why confidence is documented rather than inferred. What is not documented, and must not be invented, is that notes in these sixteen countries have already been captured. The test is still dark to other users. The filing is about the announced architecture and the markets chosen for it.

Literacy counter: treat a Community Note as a claim with a production history, not as the room speaking. Ask who can write, who can rate, whether the note cites a primary document, and whether a professional check still exists beside it. If the product page says fact-checking will be retired once notes 'work,' keep a dated copy of that page. A crowd can be real. A crowd can also be staffed. The difference is evidence, not the badge that says community.

Tells

• Product copy that pairs a crowd-note pilot with a planned sunset of named fact-checking partners
• Notes collected in a market before they are visible to that market
• High-risk electoral or authoritarian states included in an early test cohort
• Correction that has no persistent institutional owner once the partner program ends

Evidence

1. Meta Transparency Center Community Notes page listing the 16-country test and fact-check transition
2. Oversight Board 10 September 2026 co-chair statement on the Latin America test
3. IFCN 9 September 2026 statement opposing substitution of notes for fact-checking
4. AFP 12 September 2026 circulation of the Oversight Board warning

Related briefs

• Sep 12, 2026Identified
  Meta opened Community Notes testing in 16 Latin American countries as a planned replacement for third-party fact-checking
  Manufactured consensusBothsidesismDisclosure theaterAstroturfing