In a pre-registered experiment, researchers from the University of Milan Bicocca tested seven types of “nudges”—messages designed to encourage positive behavior. These nudges included reminders of social norms, prompts to cool down emotions, or messages meant to inspire empathy, deliberation, or concern for reputation. The goal was to reduce hateful speech online.

Surprisingly, none of the nudges reduced hate speech. However, they did have an unexpected benefit: they boosted engagement with harmless and wholesome content. Managing hate speech remains a major challenge for online platforms, but this finding offers a new angle to explore.

Marginal improvement

In the study, over 4,000 Americans, recruited through the platform Prolific, interacted with a replica of Facebook’s newsfeed. Participants were randomly assigned to see one of the nudges, or no message at all if they were in the control group. They then viewed 14 posts, which varied in how harmful they were, and rated how “abusive” or “hateful” the posts seemed.

While the nudges didn’t reduce engagement with harmful posts, they did lead participants to engage more with harmless ones. The difference in harmful post engagement between those who received nudges and those who didn’t was minimal.

According to the researchers, even if nudges only boost engagement with harmless content, they could still be valuable. By increasing attention to positive content, platforms might be able to crowd out harmful material.