The Effects of AI Conversations on Attitudes Towards Weather and Climate Science Conspiracies

Abstract

This study examines the effects of artificial intelligence (AI) chatbot conversations on attitudes towards weather and climate science conspiracy theories. Although prior research suggests conspiratorial beliefs are resistant to change, emerging evidence indicates AI-mediated interactions may serve as an effective tool for debiasing misinformation and conspiracy beliefs. The present experiment recruited 220 participants, who were randomly assigned to either an experimental condition, in which they engaged in a conversation with a refutational AI chatbot about their chosen weather or climate-science conspiracy, or a control condition involving nonconspiratorial discussion topics. Participants reported their belief levels and motivation to defend those beliefs both before and after the conversation. Results revealed a significant main effect of AI-conversation for reduction in conspiratorial beliefs from pre-test to post-test, and participants in the experimental condition exhibiting greater debiasing as compared to the control group. There was no significant main effect of the AI conversation for motivations to defend beliefs from pre-test to post-test as motivations increased significantly in the control condition and a non-significant decrease was found within the experimental condition. These findings suggest brief AI chatbot conversations may effectively reduce endorsement of weather and climate science conspiracy theories, even among individuals with pre-existing conspiratorial beliefs. However, changes in motivated reasoning processes appear more nuanced and require further investigation. Overall, this study extends existing literature by applying AI-based debiasing strategies to weather and climate science contexts, highlighting the potential of conversational AI as a tool for mitigating misinformation and improving public understanding of environmental risk.