Software robots initiating political discussion on Twitter are influencing the political discourse and threatening the integrity of the 2016 US presidential election, a new study has warned.
Masquerading as human beings, software robots (or social bots) on social media may be influencing public opinion as never before, according to researchers from the University of Southern California's (USC) Viterbi School of Engineering.
This could threaten the integrity of the upcoming US presidential election, the researchers warned. There are concerns that robot-generated tweets created between September 16 and October 21 could have distorted political online discussion may impact the outcome of the impending in the US.
In the study, state-of-the-art bot detection algorithms were leveraged by the researchers Emilio Ferrara, research leader at the USC, and Alessandro Bessi, visiting research assistant at USC, to analyze 20 million election-related tweets. They found that robots rather than people produced 3.8 million tweets or 19 percent of them.
Social bots also accounted for 4,00,000 of the 2.8 million individual users or nearly 15 percent of users on the microblogging website.
Researchers found that a high percentage of robot-produced tweets on Twitter indicated support for either Donald Trump or Hillary Clinton.
They found that political discussions involving the Republican candidate created by software robots were almost uniformly positive with the express purpose of distorting online discussion.
By contrast, only half of bot tweets spoke positively of his Democratic rival Hillary Clinton.
Due to the sophistication of the social bots, it is very difficult to determine who created them.