Skip to main content

A new project is trying to track hateful users’ activity on Twitter

A new project is trying to track hateful users’ activity on Twitter

Share this story

The Twitter bird logo in white against a dark background with outlined logos around it and red circles rippling out from it.
Illustration by Alex Castro / The Verge

As Twitter, like other social media companies, continues to deal with extremism on its platform, a new project is pledging to monitor how hateful users are behaving.

A newly released tool called the Exploring Online Hate dashboard is a joint project from the New America Foundation and the Anti-Defamation League, the latter of which is dedicated to tracking hate groups. The dashboard looks for trends in hateful activity through a sample of 1,000 Twitter accounts that the groups say show “hateful content directed against protected groups.” According to the project methodology, researchers selected 40 accounts that showed hateful conduct, and then algorithmically generated a larger dataset of related accounts.

Monitors 1,000 accounts

By crunching data from the Twitter accounts, the groups say, researchers can take the pulse of online hate in real time, then relay information on trending topics and sources of discussion. As of Monday afternoon, trending hashtags included both the innocuous, like #midterms2018, and the troubling, like references to the conspiracy #qanon. Top keyword terms seemed to focus on Sen. Elizabeth Warren’s DNA ancestry test. The top source referenced in the tweets was YouTube.

The dashboard looks very similar to, and shares limitations with, another project. Hamilton 68, an online tool that purports to track Russian influence operations online, has been widely cited in news reports but also criticized for failing to show its work by publicizing its secret list of monitored accounts. The online hate dashboard, maybe necessarily, also keeps its list of accounts secret, making independent examination impossible. Publicizing the list, its creators explain, could lead to changes in the listed users’ behavior.

The groups say they will also publish quarterly reports on the activity. “Our approach is designed to enable a deeper understanding of the themes, misinformation, and disinformation being disseminated by this network,” the groups explain.