Musk's election campaign: nine out of ten false X posts go uncorrected

The community fact-checking system of the X social media platform, owned by Elon Musk, failed to flag more than 90 percent of the false claims circulating about the November 3 U.S. midterm election, and the same is true of Musk's own posts. This is the finding of a study published Wednesday by NewsGuard, an organization that rates the reliability of news sources.

According to the analysis published in NewsGuard's Reality Check newsletter, researchers collected 189 posts on October 19 that spread election claims already debunked in the organization's database. Of these, by October 20 only 13 — 6.9 percent — had received a community note, while the remaining 176 posts, or 93.1 percent of the sample, remained uncorrected, even though they had racked up millions of interactions before the election. X did not respond to requests for comment from either NewsGuard or AFP.

According to the study, the false claims include fabricated and AI-generated images, slander against candidates, and baseless accusations of election fraud. One story, attributed to an alleged insider, claimed that the voting machine maker Dominion — now operating under the name Liberty Vote — bribed politicians to rig the midterm election. A conservative commentator claimed that New York and California had banned voters from showing ID and banned election officials from requesting it. No such blanket ban exists in either state, yet the post was seen by about 179,000 people and received no correction. Another post falsely portrayed a two-hour extension of voting in Dallas County, Texas, in March as fraud committed in favor of the Democratic Party. In reality, a judge ordered the extension because a local rule had confused voters, and the Texas Supreme Court suspended the ruling that same night, ordering that ballots cast after 7 p.m. be segregated. Even so, the post remained uncorrected for more than seven months and was seen by over 270,000 people. NewsGuard says that 55 percent of the appearances of the 27 false claims it tracked during this election cycle surfaced on X, and fourteen of them appeared there first — meaning the platform not only spreads election disinformation but, to a large extent, produces it itself.

The most embarrassing finding concerns Musk himself. According to NewsGuard, the owner of X — who is followed by 241.5 million people on the platform, more than anyone else — spread at least four false claims about the midterm election, and none of his posts received a community note. Musk, meanwhile, has for years been the system's biggest champion. In 2024, on Joe Rogan's podcast, he said community notes were "great" and that they "fact-check everyone, including me." Experience, however, contradicts this. According to a 2024 analysis by the Center for Countering Digital Hate (CCDH), between January 1 and October 23 of that year Musk published at least 87 posts spreading election claims that fact-checkers had rated false or misleading. These posts were viewed more than two billion times in total, and not one of them received a correction.

The community notes system began in 2021 as a pilot program called Birdwatch, and was rolled out worldwide in December 2022, six weeks after Musk bought Twitter for $44 billion. The notes are written and rated by approved users, and only appear publicly if users who typically hold differing views agree on them. According to the researchers, this condition is rarely met on partisan political issues, so the system works well when there is broader consensus — for example, on vaccine-related misinformation — but breaks down precisely on divisive election-related questions. In the United States in 2025, Meta cited X as it replaced its independent fact-checkers with its own community notes system, and is testing the system in other countries as well, even though the company's own oversight board warned in March of this year that such labels are no substitute for genuine fact-checking. NewsGuard's study, moreover, was published just as the Democratic leaders of the congressional intelligence committees warned that Russia is trying to influence the midterm election with AI-generated fake videos and ads — meaning that those spreading disinformation have at their disposal a massive platform whose verification system filters out almost nothing.

Source: AFP