Comparing button-based chatbots with webpages for presenting fact-checking results: A case study of health information

The unprecedented proliferation of online health misinformation poses a potential threat to public health. In recent times, several fact-checking organizations have adopted chatbots to present fact-checking results. However, it is unclear whether chatbots are more appropriate than traditional fact-checking websites for presenting these rigorous corrective messages. To answer this question, we compared a button-based chatbot with a traditional webpage for presenting fact-checking results. As fact-checkers’ expertise cues could influence users’ perception of fact-checking, we also considered the effect of expertise cues in our study. We conducted a 2 (interaction type: webpage vs. chatbot) × 2 (expertise cue: non-highlighted vs. highlighted) between-subjects online experiment (N=308). The results show that the chatbot leads to higher perceived ease of use, which in turn increases the effectiveness of fact-checking. The highlighted expertise cue tends to decrease users’ intention to use, especially when they interact with the webpage. Finally, we discuss the feasibility of using chatbots to disseminate fact-checking content and several design implications for the creation of an effective tool to fact-check health information.

Duke Scholars

Author Yucheng Jin DKU Faculty