Tackling harmful online comments with targeted email feedback

Platforms that host online discourse face competing demands between allowing free speech and enforcing community guidelines meant to encourage users’ safe participation. One tool they have is content moderation, much of which is done by deleting content without providing feedback to or even notifying the offending user, making it hard for users to understand what they did wrong. In a joint project between ETH Zürich and the Swiss newspaper Blick, we test whether providing feedback after deletion affects users’ future commenting behavior. We identify toxic comments that were submitted but not published on Blick.ch and compare the default lack of feedback to three different email responses: a generic response informing users that their comment violated the guidelines, an empathy-based response encouraging them to think of the targets of their offensive content, and a norms-based response highlighting the low toxicity of other comments. Initial results suggest that feedback does indeed improve users’ future comments.

Diese Session ist Teil des Themenschwerpunkts "Freiheit" im Rahmen des Wissenschaftsjahres 2024.