Feb. 5 (UPI) -- More than 20 percent of political Twitter posts leading up to the midterm elections last fall were automated, or "bot" posts, from Russia and Iran, a new analysis has found.
Researchers at the University of Southern California identified more than 15,000 bot accounts using the #ivoted hashtag on Election Day, the study released Tuesday by USC's Information Sciences Institute and a Swiss researcher found.
The details provide more insight into how some are trying to influence the U.S. political process.
Researchers said the phony accounts amplified political divisions and posted stories meant to inflame and trigger emotional responses.
"We have the more liberal bots that are more inflammatory, they create more of this chaos and confusion if you like, and on the other hand, however, the conservative bots are more influential," said Emilio Ferrara, the author of the study. "Conservative bots have a much more prominent position in these information sharing networks. They project a stronger influence on the human users."
Thousands of the bots were traced to Russia and hundreds to Iran, Ferrara said. Other accounts seemed to originate from the United States.
The study identified the bots using a sophisticated tool developed at the University of Indiana called a "Botometer," which tries to determine whether an account is operated by a human or a robot. Researchers say it's accurate about 95 percent of the time.
Twitter said last October it deleted 9 millions accounts in the run-up to Election Day, but added there were others that weren't discovered. Twitter said it removed the "vast majority" that appeared to be tied to Russia.