Scaling crowdsourcing interventions to combat partisan misinformation

Partisan misinformation hijacks people’s right to make decisions based on accurate information, posing a threat to democracy and liberal values. Current interventions to counter misinformation are less effective when it comes to politically polarizing content, especially among extreme partisans who share the most misinformation. A new line of research suggests that crowdsourcing interventions, or using laypeople’s judgments to help people spot misinformation, provide an additional layer of content moderation that can help overcome these limitations. We present a model that explains when crowdsourcing interventions will be successful based on three factors: trust in fact-checking sources, dissonance with previous beliefs, and crowd size. These three factors are often at odds in politically polarized social media environments, leading to less effective fact-checking. Based on this model, we discuss how crowdsourcing interventions could be scaled in a way that is ethical and leverages artificial intelligence to connect people with neighboring communities outside their ideological echo chambers. Finally, we propose venues for future research in the field of crowdsourcing interventions that lie at the intersection between individual-level and system-level solutions to partisan misinformation.