NEW YORK (Reuters) - Fake accounts posting about the U.S. presidential election are proliferating on the social media platform X, according to a social media analysis company's report shared with Reuters exclusively ahead of its release on Friday.

Analysts from Israeli tech company Cyabra, which uses a subset of artificial intelligence called machine learning to identify fake accounts, found that 15% of X accounts praising former President Donald Trump and criticizing President Joe Biden are fake. The report also found that 7% of accounts praising Biden, a Democrat, and criticizing Trump, a Republican, are fake.

Cyabra's study is based on a review of posts on the X platform, formerly known as Twitter, over two months beginning March 1. The review included analyzing popular hashtags and determining sentiment in terms of whether posts are positive, negative or neutral.

The analysis shows that newly detected fake accounts had increased up to tenfold during March and April.

The report cites 12,391 inauthentic pro-Trump profiles out of 94,363 total and 803 inauthentic pro-Biden profiles out of 10,065 total.

A spokesperson for X did not respond to a request for comment about the fake accounts, nor did representatives from the White House and Trump campaign.

X and other social media platforms have been under greater scrutiny since 2016, when Russia interfered in the U.S. presidential election in an attempt to boost Trump's candidacy and harm his opponent, Democrat Hillary Clinton. Election officials and online misinformation experts are again watching for misleading narratives ahead of the Nov. 5 election.

The fake accounts praising Trump this cycle are part of a coordinated campaign to sway public opinion and influence online discussions, Cyabra said. The report did not identify the individuals or groups behind the campaign.

Cyabra said it made that determination based on evidence including the use of identical hashtags and the fact that fake accounts published posts and comments at the same time. The report found that the fake pro-Trump accounts pushed two main messages: "Vote for Trump" and "Biden is the worst president the U.S. has ever had."

"The level of coordination suggests that there is a nefarious objective and that there is a whole operation in order to change people's opinion," said Cyabra's vice president, Rafi Mendelsohn.

The fake accounts backing Biden are not part of a coordinated campaign, the report said, as the hallmarks of a coordinated campaign - such as fake accounts posting at the same time - were not identified.

X, which was publicly held until its 2022 takeover by billionaire Elon Musk, has long downplayed the use of fake accounts on its platform. Twitter said in May 2022 that fewer than 5% of its daily active users were "false or spam" based on an internal review of accounts. At the time, Cyabra had estimated that 13.7% of Twitter profiles were inauthentic.

In an X post on April 4, Musk wrote that a "system purge of bots and trolls" was under way and that the company "will be tracing the people responsible and bringing the full force of the law to bear upon them." In October the company tested its "Not a Bot" program in New Zealand and the Philippines to combat bots and spammers.

(Reporting by Stephanie Burnett in Amsterdam and Helen Coster in New York; Editing by Colleen Jenkins and Matthew Lewis)

By Stephanie Burnett and Helen Coster