(Reuters) – The crowd-sourced fact-checking feature of Elon Musk’s X, Community Notes, is “failing to counter false” claims about the U.S. election, the Center for Countering Digital Hate (CCDH) said in a report on Wednesday.

Out of the 283 misleading posts that CCDH has analyzed, 209 or 74% of the posts did not show accurate notes to all X users correcting false and misleading claims about the elections, the report said.

“The 209 misleading posts in our sample that did not display available Community Notes to all users have amassed 2.2 billion views,” CCDH said, urging the company to invest in safety and transparency.

X did not immediately respond to a Reuters request for comment.

X launched its “Community Notes” feature last year, which allows users to comment on posts to flag false or misleading content, in effect crowd-sourcing fact checking to users rather than a dedicated team of fact checkers.

The report comes after X lost a lawsuit brought by CCDH earlier this year that faulted it for allowing a rise in hate speech on the social media platform.

Social media platforms, including X, have been under scrutiny for years over the spread of misinformation and conspiracy theories, including false information about elections and vaccines.

Last month, secretaries of state from five U.S. states urged billionaire Musk last month to fix X’s AI chatbot, saying it had spread misinformation related to the Nov. 5 election.

Musk, who last month endorsed Republican presidential candidate Donald Trump, himself has been accused of spreading misinformation. Polls show Trump is in a tight race with Democratic Vice President Kamala Harris.