Project description

This PhD project studies the behaviors of social bots, i.e., social media accounts controlled by software or algorithms rather than humans, in online science communication, especially their interactions with human accounts, and the effect of these behaviors: what kind of content are social bots more likely to (re-)post? How and to what extent does social bot activity influence the public perception of science? And how could human users detect social bots to avoid their influence?

This project particularly plans to investigate two aspects of social bot behavior: the content social bots tend to post and share and the interaction between social bots and human users. Social media content sent and reposted by both social bots and human accounts will be compared by topic modelling to show topic and rhetoric differences, which demonstrates the themes that social bots attempt to promote, e.g., whether social bots are more likely to post anti-science or pseudoscience content. Next, based on network analysis, human accounts which are more likely to interact with social bots are to be identified, which could provide potential guidance to help the public detect bots.

Further, the potential effect of content driven by social bots on human users and their capacity to detect social bots are to be examined under experimental settings. What kind of factors could influence a human’s capability to identify social bots? How could this capability further help people detect bot-driven misinformation? What kind of intervention could be applied to effectively debunk such misinformation? The aim of this study is to produce insights into social bot behavior in science communication, to better facilitate the process of public understanding of science, and debunk scientific misinformation in social media.

Cooperations

Prof. Dr. Guido Zurstiege (University of Tübingen, Institute of Media Studies)