Bot or Not: A Twitter Bot Detection Tool

You might not know all of your Twitter followers, but you probably expect them to be human. That’s not what new research suggests.

A recent study conducted by the University of Southern California and Indiana University says that up to 15 percent of Twitter users are actually fake “bots.”

Some of these accounts are beneficial, like those that alert people about earthquakes.

But bots also have a downside — and can play a role in online harassment and disseminating propaganda and misinformation.

Researchers working on the Truthy project (a project aimed at studying information diffusion in social media) at Indiana University hope to make it easier to find these bots. They created a new tool called Bot or Not for researchers and reporters.

“We see this task as a new facet of media literacy,” wrote Clayton A. Davis, one of the developers, in an email. “Traditional media literacy teaches people to be skeptical of the content they read on the web; BotOrNot hopefully illustrates that one should be skeptical of the source, specifically the possibility that the content is being promoted by a bot.”

Users can type in a screen name and see an account’s “bot score.” The algorithm uses several features to determine a score, including hashtag patterns, language, timing of posts, post language, followers and mentions. The higher the score, the more likely the account is a “bot.”

We put it to the test, with mixed results. You can see more details in the video, including how to figure out your own “bot score.”

All semester, the Annenberg Truth Squad has been looking at problems faced by journalists, the news industry and audience trust. Now, we’re looking at solutions — web sites, apps and new initiatives.