Processing Fabricated Foreign Risks with AI Fact-Checker

The role of Third-Person Perception, Nationalism and Conspiratorial Orientation

Abstract

The proliferation of fake information has become a global challenge, with conspiracy content particularly prone to misleading the public, making it a key focus of risk communication governance. The widespread application of AI in news writing and verification has transformed news interaction, with articles increasingly labeled as “verified by a human” or “verified by AI,” creating distinct interfaces for news evaluation. Individual personality traits, such as nationalism and conspiracy orientation, are crucial in shaping responses to risk information, while group influences increasingly guide opinions and behaviors. The third-person effect theory provides a valuable framework for understanding how people perceive and react to information based on others’ opinions. This study explores how personality traits and verification interfaces influence users’ information judgment and behavior through the third-person effect. A group experiment randomly assigned participants to a human-checker group (N=254) or an AI-checker group (N=252). Participants read news labeled with different verifiers and completed questionnaires assessing their perceptions of others’ opinions and their willingness to seek information. The study examines how personality traits interact with verification interfaces to affect information behavior, offering theoretical insights for optimizing verification strategies. Results indicate that verification interfaces significantly shape the relationship between personality traits and the third-person effect, influencing information-seeking behavior. In the human verification interface, blind patriotism positively predicts perceived social consensus (β = 0.257, p = 0.001), which indirectly increases information-seeking behavior (indirect effect = 0.032), suggesting that individuals with strong blind patriotism are more motivated to seek information based on social consensus. In contrast, conspiracy preference has limited influence on information-seeking in this interface (β = 0.123, p = 0.006). In the AI verification interface, blind patriotism continues to positively predict social consensus (β = 0.216, p = 0.004), though its indirect effect on information-seeking is slightly stronger (indirect effect = 0.046), indicating a weakened link between patriotism and information-seeking. Notably, conspiracy preference significantly predicts information-seeking behavior (β = 0.168, p = 0.032) and has a marginally significant effect on third-person perception (β = -0.147, p = 0.119). This suggests that AI-verified interfaces trigger the exploratory motivation of high-conspiracy individuals, heightening their attention to others’ reactions and promoting active information seeking. Overall, the type of verification interface moderates the effect of personality traits on third-person perceptions and information behavior, with the third-person effect serving as a central mediator. These findings deepen understanding of how personality and interface design shape public engagement with verified information, offering theoretical guidance for improving information verification strategies and enhancing media literacy.