X’s crowd-sourced fact-checking system, Community Notes, is under renewed scrutiny after a report by media credibility watchdog NewsGuard found that it failed to flag 93.1 per cent of the false claims about the US midterm elections examined in its analysis. What is most concerning about the results of this study is that community-led moderation rarely can solve political misinformation problems when it spreads quickly on social media platforms.
According to NewsGuard, only 13 of the posts with false election-related claims identified in its analysis received Community Notes. The rest were still circulating and drew millions of interactions from people looking to see how they got their news during the November 3 U.S. midterm elections. This shows how misinformation can be confused with accurate information.
The results also call attention to Elon Musk, who has repeatedly praised Community Notes as an alternative to traditional fact-checking programs. NewsGuard found at least four false claims about the midterm elections promoted by Musk but none received a Community Note, the watchdog said.
The study has reignited the debate on whether social media platforms can depend on users to identify and correct misleading content without professional fact-checkers. Community involvement can add context to the false claims and the NewsGuard results show that a significant number of false claims may go unaddressed.
False voter identification claims highlight moderation challenges
One of the examples cited by NewsGuard was a conservative commentator who claimed that New York and California had prohibited voters from displaying identification and election officials from asking to see it. The watchdog called that claim false as neither state had enacted such a rule.
Voting procedures may be misleading to voters and misleading to voters about identification requirements, ballot access or election administration. When it is common, the claims spread widely without correction, users might find it hard to distinguish between true information and unsubstantiated claims.
False claims of electoral fraud can also undermine public confidence in democratic institutions. They could be amplified when they are shared by prominent accounts or when there is a large amount of engagement before a correction becomes available.
The NewsGuard analysis illustrates how a community-based moderation system may struggle to address these challenges consistently. Even if a claim can be checked against publicly available information, a correction may not appear if contributors fail to reach the agreement required by the platform's system.
How X’s Community Notes system works
Community Notes was launched on X, formerly known as Twitter, in late 2022. That feature permits eligible contributors to write explanations which add some context or correct potentially misleading posts. Unlike traditional fact-checking programmes it is based on user-generated content and ratings.
To get the note to be seen by a wider audience the system is designed to have agreement among contributors who have at different times disagreed on other notes. This will ensure that partisan preferences are less influential and that corrections are more credible to those who are not Democrats.
Musk has publicly supported the feature and has said it is an effective way to give context to posts -even those that are posted by the most influential users. Proponents of the model say a distributed system can help identify misinformation on a vast platform without having a central team of fact-checkers.
But the same consensus-based design can be problematic when political claims are highly divisive. If contributors cannot agree that a note is helpful, a correction may not appear even when a post contains a demonstrably false statement.
NewsGuard's findings do not indicate why each individual post went uncorrected. Nor does the 93.1 per cent figure alone indicate how Community Notes works on every election-related post on X. Nevertheless, the results suggest there is a substantial weakness in the particular set of claims studied by the watchdog.
Research presents a mixed picture of crowd-sourced fact-checking
Moderation of the community has been mixed, as studies on community-driven moderation have reported. Such mechanisms can help to add context to misleading content and have been studied to combat misinformation on such topics as vaccines. But their effectiveness may also be dependent on the content of the claim, the availability of sufficient evidence for the claim, and the ability of those with different points of view to reach the same conclusion.
Political misinformation is a particularly difficult test as elections often involve strong ideological loyalties. Contributors may disagree not only on the interpretation of events but also about which sources they consider trustworthy.
A consensus-based model can help guard against corrections driven by a single political group. But it can be slower or less effective when contributors are divided over contentious issues. The NewsGuard study adds to the questions facing platforms that increasingly rely on community participation to moderate information at scale.
Meta's move towards community-led moderation
The X debate is also relevant to broader technology industry developments. Meta, the parent company of Facebook and Instagram, announced the end of its third-party fact-checking programme in the US in January 2025 and said that it would move towards a Community Notes-style approach.
Meta CEO Mark Zuckerberg stated that the company’s previous system had led to excessive censorship and this was disputed by professional fact-checking organisations. Meta began testing its own community-based feature in 16 Spanish-speaking countries in Latin America, according to the information reported in connection with the initiative.
This is part of a wider rethinking of the way technology companies balance freedom of expression, content moderation and how to correct for false information. Community-led moderation is not a substitute for professional fact-checking, particularly when claims concern elections and other matters of public importance.
Meta's Oversight Board has also raised questions about whether a Community Notes-style system can replace professional fact-checking. Such models will work better if they provide timely, correct and visible corrections rather than just letting users write notes.
What the findings mean for social media users
The NewsGuard study underscores a fundamental challenge for platforms that rely on crowdsourced moderation: Even when users can participate in fact-checking, it is not guaranteed false claims will be corrected. When an entire system is designed to ensure that people are able to agree across some political lines, some misleading posts will still spread without any context.
For users, the findings underline the importance of checking election-related statements against official election authorities and multiple credible news sources. A post without a Community Note should not be taken to be accurate, as can be argued, as the presence of a note should not substitute for independent verification.
As social media companies experiment with community-based moderation, their success will depend on whether those systems can spot false stories and respond in time so that they can be corrected quickly enough to reach the target audience. The NewsGuard findings put that challenge more firmly in focus, especially as platforms are working to manage political information diffusion during elections.
Comments
Leave a Comment