Empathy-Based Counterspeech Can Reduce Racist Hate Speech in a Social Media Field Experiment

PNAS December 14, 2021 118 (50) e2116310118; https://doi.org/10.1073/pnas.2116310118

Commentary

Despite growing awareness of the negative impact of hate speech on social media platforms, there is little consensus on approaches to reduce it.

Content moderation by governments and social media companies can curb online hostility, but such measures can suppress malicious as well as valuable speech, which may differ from the goal of reducing malicious speech.

An alternative strategy is to employ counterspeech.

There is little experimental evidence on the effectiveness and design of counterspeech strategies, even though this is gaining popularity as agencies and others adopt it.

Modeling an intervention on current I / NGO practice, a message containing xenophobic (or racist) hate speech was sent, and English-speaking Twitter users were randomly assigned to one of three counterspeech strategies or a control group: empathy, consequence warning, and humor.

According to an ITT analysis of 1,350 Twitter users, empathy-based hate speech messages could increase the retroactive deletion of xenophobic hate speech by 0.2 SD and decrease xenophobic hate speech by 0.1 during a four-week follow-up period. However, the strategy of using humor and consequence warnings has no consistent effect.