Newswise — A surprisingly high percentage of the political discussion taking place on Twitter was created by pro-Donald Trump and pro-Hillary Clinton software robots, or social bots, according to research published by First Monday, an online journal published by the University of Illinois at Chicago Library.

The researchers, who are computer scientists at the University of Southern California’s Viterbi School of Engineering, worry that robot-generated tweets distort political online discussion and may impact election outcomes.

“Software robots masquerading as humans are influencing the political discourse on social media as never before and could threaten the very integrity of the 2016 U.S. presidential election,” said Emilio Ferrara, research leader at the USC Information Sciences Institute and research assistant professor of engineering.

By leveraging state-of-the art bot detection algorithms, Ferrara and study co-author Alessandro Bessi, USC visiting research assistant at the institute, analyzed 20 million election-related tweets created between Sept. 16 and Oct. 21. They found that robots, rather than people, produced 3.8 million tweets, or 19 percent. Social bots also accounted for 400,000 of the 2.8 million individual users, or nearly 15 percent of the population under study.

Their recently released paper titled, “Social Bots Distort the 2016 U.S. Presidential Election Online Discussion,” highlights three ways in which the presence of these bots can affect the dynamics of political discussion:

Influence can be redistributed across suspicious accounts that may be operated with malicious purposes. The political conversation can become further polarized. Spreading of misinformation and unverified information can be enhanced.

Interestingly, the researchers found Trump’s robot-produced tweets were almost uniformly positive, boosting the candidate’s popularity. By contrast, only half of Clinton bot tweets were positive, with the other half criticizing the nominee. South Carolina produced the most fake campaign-related tweets, the study reports.

Because of social bots’ sophistication, it’s often impossible to determine who creates them, although political parties, local, national and foreign governments and “even single individuals with adequate resources could obtain the operational capabilities and technical tools to deploy armies of social bots and affect the directions of online political conversation,” the report says.

The “master puppeteers” behind influence bots, often create fake Twitter and Facebook profiles, according to Ferrara and Bessi.

“They do so by stealing online pictures, giving them fictitious names, and cloning biographical information from existing accounts,” the researchers wrote in their study. “These bots have become so sophisticated that they can tweet, retweet, share content, comment on posts, ‘like’ candidates, grow their social influence by following legit human accounts and even engage in human-like conversations.”First Monday is one of the first openly accessible, peer–reviewed journals solely devoted to research about the Internet. The journal is produced by the UIC Library’s electronic journals publishing program.

For additional information about the research, contact Amy Blumenthal, media relations for USC Viterbi School of Engineering, [email protected]; (213) 821-1887.