Botometer

Website

Botometer

Founded in 2014

Botometer is a web-based program that uses machine learning to classify Twitter accounts as bot or human by looking at features of a profile including friends, social network structure, temporal activity, language and sentiment. Botometer outputs an overall bot score (0-5) along with several other scores that provides a measure of the likelihood that the account is a bot.

Botometer is not affiliated with RAND. It was selected for this database because it fits our researchers' inclusion criteria.

Tool type

Bot/spam detection

Status

Fully operational

Intended users

General public

Cost

Free

Tool focus

This tool is process-focused. It evaluates how information is produced and disseminated.

Method or technology

Machine learning and AI

Is the tool automated?

Yes

Founding organization

Indiana University

Founder/primary contact

Clayton Davis, Onur Varol

How is this tool working to address disinformation?

This tool aims to fight disinformation by identifying bots (frequent sources of disinformation) and their content, ultimately helping users interpret and evaluate information.

Is there a connection with tech platforms?

None found

Who is funding the tool?

Botometer is a collaboration between Indiana University Network Science Institute (IUNI) and the Center for Complex Networks and Systems Research (CNetS).

Are there external evaluations?

A 2017 evaluation by the tool developers focuses on the tool's abilty to detect bots, rather than the value of bot detection in fighting disinformation: arxiv.org