Over the past three years, my research project on the connection between online speech and offline harm has achieved major milestones. We have been able to legally and ethically acquire millions of social media posts from India for statistical analysis and have conducted large-scale surveys. I will describe our results in more detail once the corresponding papers have passed peer review.

The most important insight is this: In sufficiently divided and politicized environments, even seemingly harmless online speech can provoke conflict-related emotions such as anger and fear, and affect overall sentiments towards outgroups. Additionally, such seemingly harmless online speech is also a leading indicator of physical attacks. This situation makes it very difficult to devise effective policy responses.

Living with such negative effects of social media technologies is problematic, but censorship of legally protected and non-violent speech carries its own set of risks and disadvantages, especially in Democracies. A much more nuanced and open debate will be needed to find meaningful solutions in this area.