The newest tool in the Observatory on Social Media’s (OSoMe) toolbox has arrived. The research center just launched an open beta version of BotSlayer for the public to try.
BotSlayer, which expands off of the OSoMe tool Bot Electioneering Volume (BEV), can identify bot campaigns on Twitter before they escalate.
But while BEV was created specifically to visualize bot activity on Twitter during the 2018 U.S. Midterm Election, BotSlayer will run continuously.
OSoMe Director Filippo Menczer said applications like BotSlayer are important in preventing bots from manipulating public opinion.
“We developed BotSlayer to make it easier for journalists and political campaigns to monitor potential new disinformation campaigns that attempt to manipulate public opinion using bots,” Menczer told IU News.
BotSlayer works by identifying tweets that have cooccurrences among likely bots. The tool then flags hashtags, links, media, and accounts—referred to as entities—that appear to be a part of coordinated bot campaigns and gives those entities a score. The higher the “BS” score for an entity, the more likely the entity is part of a bot-led agenda.
BotSlayer also depends on other tools in the OSoMe portfolio, which IUNI has had an important role in developing.
One key component of BotSlayer’s work is correctly identifying bots, which it uses Botometer to accomplish.
BotSlayer also uses Hoaxy to help users visualize their BotSlayer results. After receiving the BS Level for an entity, users can click a Hoaxy button to watch how that campaign spread, identify who in the campaign is a bot and who is not, and see a timeline for the campaign’s tweets.
IUNI Associate Full Stack Developer Marc McCarty built BotSlayer’s front end, working closely with Pik-Mai Hui and Kai-Cheng Yang at the Center for Complex Networks and Systems Research (CNetS) on the tool.
The app, which has been in closed alpha, arrived just in time for the public to test during the 2020 Democratic Party presidential debate on Sept. 12—which could come in handy. McCarty says previous Democratic Party presidential debates actually inspired additional BotSlayer features.
In the first debate, for example, bots began a campaign targeting Kamala Harris by pushing the agenda that she is not black because she is half Jamaican and half Indian.
That campaign spurred the feature that allows BotSlayer to remember text used in hashtags, usernames, and media links that the app has flagged as part of a campaign. So if the Kamala Harris bot campaign occurred again today, users might find text tweeted in the campaign, including the words “Jamaican” or “Indian,” trending on BotSlayer with a high BS score surrounding the text.
If you’re interested in trying BotSlayer for the debate tonight or for everyday use, check out this page on how to begin the process. You can learn about the larger truth-seeking mission of OSoMe by visiting truth.iu.edu and by watching the video below.