This dissertation explores the impact of bots on online public discourse, specifically focusing on human users’ language and attitudes in response to online interactions. The findings from Study 1 suggest that while bots generally exhibited lower levels of politicization, polarization, and neutrality, they displayed higher levels of anger, disgust, fear, and joy. Some of these features, specifically, politicization and disgust of bots can still influence humans over time. Furthermore, Studies 2 and 3 compared the effects of LLM-generated content versus non-LLM-generated content on individuals’ attitudes. The results show that LLM-generated contents can subtly influence users, as individuals often struggle to distinguish between human-like and machine-generated content on social media. As bots become more sophisticated with technological advancements, they are increasingly capable of shaping human attitudes in ways that are nearly indistinguishable from human interactions. Given the pervasive use of such technologies on social media, understanding their relational impact on humans is becoming crucial.
Read
