X’s Community Notes Failed To Flag 93% Of False US Midterm Claims, Study Finds

A crowdsourced fact-checking system on Elon Musk’s social media platform X failed to flag 93.1 per cent of false claims related to the US midterm elections, including allegations promoted by Musk himself, according to a study by US media watchdog NewsGuard.

The study, released on Wednesday, found that only 13 of dozens of posts containing false election-related claims had been flagged by Community Notes, X’s system that allows users to add context or corrections to posts. The uncorrected posts had attracted millions of interactions ahead of the November 3 midterm elections, NewsGuard said.

Musk’s Posts Among Those Not Flagged

Community Notes was introduced on X in late 2022 and is increasingly being adopted as a model by other technology platforms. The system relies on contributions from users to provide additional context or correct misleading information.

NewsGuard identified dozens of posts spreading false claims about the US midterm elections, including allegations of election fraud. Among them was a post by a conservative commentator claiming that New York and California had made it illegal for voters to show their voter identification and for election officials to ask for it. No such rule exists in either state, the watchdog said.

Musk, who has frequently praised Community Notes as an effective tool against misinformation, has described the system as “awesome” and argued that it allows users, including himself, to be corrected by the crowd.

However, NewsGuard said Musk, who has 241.5 million followers on X, had promoted at least four false claims related to the midterm elections. None of those posts carried a Community Note.

“X’s crowdsourced fact-checking program failed to flag 93.1 percent of false claims related to the US midterm elections,” NewsGuard said.

X did not respond to Agence France-Presse’s (AFP) request for comment.

Questions Over Crowdsourced Fact-Checking

Researchers have repeatedly questioned the effectiveness of Community Notes in countering misinformation, particularly on politically polarised subjects. While studies have found that the system can help dispel some falsehoods, including vaccine misinformation, researchers have cautioned that it works best when there is broad agreement on the facts.

Some have also warned that contributors may be influenced by partisan beliefs and could use the system to target political opponents.

Technology companies have increasingly looked to community-driven moderation as an alternative to professional fact-checking. Last year, Meta ended its third-party fact-checking programme in the United States, with CEO Mark Zuckerberg arguing that it had resulted in “too much censorship”. Professional fact-checkers have strongly rejected that claim.

Last month, Meta said it had begun testing its own version of Community Notes in 16 Spanish-speaking countries in Latin America as part of its move towards replacing its fact-checking programme. The company has not announced a timeline for making the change official.

Meta’s Oversight Board, a quasi-independent body funded by the company to review its content moderation decisions, has said Community Notes is not an adequate replacement for professional fact-checking.

AFP, which reported the findings, currently works in 26 languages with Meta’s global fact-checking programme, including across Latin America.