Friend or enemy? How perceived present and future influence of AI on self vs. others shape human-AI relations across personality mindsets

This study applies the third-person perception (TPP) framework to examine how individuals develop relational orientation toward AI as either a friend or an enemy. Using a national survey in China (N = 1070), we examine how perceived differences between AI's impacts on others versus oneself across both valence (positive/negative) and temporal (present/future) dimensions serve as psychological mechanisms linking media exposure and user experience with relational orientation. Findings show that both media attention and direct AI use are positively associated with perceiving AI as a friend and negatively associated with viewing it as an enemy. Media attention was linked to stronger perceptions that AI currently benefits others more than oneself, while AI use was associated with reduced perceptions that others are more negatively affected. Interestingly, perceiving that AI harms others more than oneself, both now and in the future, was positively related to viewing AI as a friend. Additionally, perceiving that AI will benefit others more in the future was also associated with affiliative perceptions. Multigroup analysis revealed that these pathways differed by implicit theories of personality: entity theorists were more influenced by future-oriented perceptions, whereas incremental theorists responded more strongly to media attention.