AI-powered bot network spreads contradictory political messages on X
Researchers discovered a network of over 400 AI-powered X accounts designed to praise Trump administration figures. The bots' messaging became contradictory when Epstein-related documents were withheld, with some accounts defending the decision while others demanded resignations, revealing the automated nature of the coordinated inauthentic behavior.
What happened
Researchers from Alethea and Clemson University's Media Forensics Hub uncovered a network of more than 400 AI-powered X accounts originally designed to post positive messages about Trump administration figures including Health and Human Services Secretary Robert F. Kennedy Jr. and White House Press Secretary Karoline Leavitt.
When Attorney General Pam Bondi announced she would not release additional Jeffrey Epstein-related documents, the bots' messaging fractured, with some accounts defending her and others demanding her resignation or that of figures like FBI Director Kash Patel and Dan Bongino within the same minute.
Reported impact
The accounts were designed to contribute to a partisan echo chamber and 'massage perceptions' according to Darren Linvill, director of Clemson University's Media Forensics Hub, though they typically had only a few dozen followers and their posts rarely received many views.
Index analysis
All accounts were created seemingly in batches around three specific days in 2024. They frequently punctuate posts with irrelevant hashtags, post almost exclusively by replying to other users, and sometimes repeat posts verbatim.
It is unclear who is behind the network or which publicly accessible AI chatbot was used to power it.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.