We study dangerous speech and ways to counteract it.

People don’t commit violence against other groups - or even condone it - spontaneously. First they must be taught to see other people as pests, vermin, aliens, or threats. Malicious leaders often use the same types of rhetoric to do this, in myriad cultures, languages, countries, and historical periods. We call this Dangerous Speech. Violence might be prevented by making it less abundant or less convincing. We work to find the best ways to do this – while protecting freedom of expression.


Online harassment resource guide

This Wikimedia resource compiles interdisciplinary research on understanding and responding to the problem of online harassment, and serves as an excellent starting point for scholars and practitioners alike.

Professor Chimène Keitner examines dangerous speech, ‘deep fakes’ in conversation

Watch (or listen) to a full conversation with DSP's Susan Benesch, Bobby Chesney, and moderator Chimène Keitner about Dangerous Speech and "deep fake" inauthentic content online.

Blog + News


Myanmar's genocide of Rohingya Muslims was shaped by uniquely modern forces, like social media and rising ultra-nationalism. But if we learn from the factors which fueled this 21st-century atrocity, we can better predict and prevent the next ones.

Wired: Twitter Releases New Policy on ‘Dehumanizing Speech’

From Louise Matsakis at Wired: Twitter will soon enact a conduct policy on dehumanization - but DSP's Susan Benesch explains the difficulty in moderating such content.