Computer science researchers at the University of Southern California created sophisticated algorithms to figure out, during a one-month period including presidential debates, how many tweets about Donald Trump and Hillary Clinton were generated not by humans, but by software.
“Looking at the content itself, but also the timing of the activity, the network structures, how information spreads from one account to another and so forth. Eventually we realized that about 15 percent of the accounts on the Twitter feed discussing elections in the U.S. are bots,” Research assistant professor Emilio Ferrara told KCBS.
More importantly, Ferrara says, these bots generated about 20-percent of the entire conversation.
Bots don’t arise spontaneously. People program them.
The research team found that the bots designed to promote Donald Trump made it appear he has grassroots support, while the bots related to Hillary Clinton have been more neutral.