A network of AI-powered bot accounts is quietly flooding X, with pro Trump messages—and it just hit a weird glitch that’s making people take notice.
Researchers say hundreds of these fake accounts are reportedly posing as conservative users, automatically replying to posts from verified X users with glowing support for Trump officials. The twist? The bots seem to have short-circuited over the controversy surrounding the Epstein files, exposing the artificial intelligence behind the accounts.
NBC News first reported the story, citing research from Alethea, a social media analytics firm, and Clemson University’s Media Forensics Hub. So far, over 400 accounts have been identified, but the researchers believe there could be many more.
These bots have been spreading praise for Trump-era figures like Health Secretary Robert F. Kennedy Jr. and press secretary Karoline Leavitt. But the point doesn’t seem to be going viral. Most of these accounts have barely any followers, and their posts rarely get much traction.
“They’re not really there to get engagement. They’re there to just be occasionally seen in those replies,” said Darren Linvill, director of Clemson’s Media Forensics Hub. He told NBC News the strategy is more about creating a steady drip of pro Trump sentiment than building big online followings.
The accounts follow some very specific patterns. Researchers say they were created in batches on just three separate days last year. They love hashtags—often ones that have nothing to do with the topic—and almost always post by replying to other users. Sometimes they even just repeat someone’s post back to them.
No one knows exactly who is behind the network, or which AI chatbot is powering it. But it’s been active since at least early 2024, steadily promoting MAGA content before and after the election.
That is, until the Epstein files controversy threw everything into chaos.
Attorney General Pam Bondi recently said she wouldn’t release additional files about Jeffrey Epstein, the late sex offender with ties to powerful figures, including Trump. That announcement caused a rift in the MAGA world—and the bots picked up on it, but not in a coordinated way.
Related: Pam Bondi Tells Trump She’s Ready to Unseal Epstein Grand Jury Files
Some accounts defended Bondi and Trump. Others went in hard, calling for Bondi or FBI Director Kash Patel and Deputy Director Dan Bongino to resign. At one point, a single account replied to two different users in the same minute with totally opposite takes on the issue.
One post declared Bondi had been cleared by the DOJ and that “no Epstein client list” was found. The same account later told people to revolt against the Trump administration, writing, “Retweet if you believe that Trump & his cronies are lying to the public and treating us like we’re stupid.”
C. Shawn Eib, head of investigations at Alethea, said this kind of split response might actually reflect the real confusion in Trump’s base. “This split reaction mimics the organic reaction among supporters of Trump’s second administration,” Eib said.
The bigger concern is how AI bots are making these kinds of campaigns easier than ever. AI can now generate realistic posts 24/7, and social platforms like X have pulled back on content moderation, especially after Elon Musk cut much of the trust and safety team when he bought the site in 2022.
The White House had no comment on the report, and neither the Department of Health and Human Services nor X responded to requests for comment.
Researchers say it’s likely there are many more of these AI-driven accounts still operating on X. But with access to platform data becoming harder for researchers to get, there’s no way to know just how big the network really is.