Teaching Machines to Understand Hate

2017 Impact Report

Online hate speech intimidates its victims and stifles public discourse. That’s why ADL is so focused on combating cyber-hate. Accurately identifying hateful content from amongst the billions of messages posted each day is a daunting challenge. ADL's CTS and University of California, Berkeley’s D-Lab launched the Online Hate Index (OHI) in 2017 to take on this challenge. 

The OHI combines artificial intelligence and machine learning with social science to uncover and identify trends and patterns in hate speech. By teaching machines to recognize hate, the OHI offers the promise of making online communities more respectful and inclusive. 

Most common words found in hate and non-hate posts
Words strongly associated with hate speech

Read more about this year's Online Hate Index report.

“The goal of the Online Hate Index is to better understand the growing amount of hate on social media and to use that information to address the problem.” - Brittan Heller, Director, The Center for Technology and Society Share via Twitter Share via Facebook

ADL in the Media