Human-AI cooperation: Modes and their effects on attitudes

Since advancements in artificial intelligence (AI) have allowed AI to live and work with humans as independent agents, it has become increasingly common for people to cooperate with autonomous AI agents, which is known as human-AI cooperation (HAIC). The purpose of this research is to explore the various modes of HAIC and their effects on people’s attitudes. The current paper presents a between-group study conducted by dividing samples into two groups, i.e., a human–human cooperation group and a HAIC group. A total of 1367 samples are investigated. By employing confirmatory factor analysis, we identify-four modes of HAIC: interdependent exploitation, independent exploitation, interdependent exploration, and independent exploration. By conducting a multigroup comparative analysis, we find that people prefer interdependent exploration modes and dislike independent exploitation and independent exploration modes when cooperating with AI. Our study provides insights into the modes of HAIC and the effects of these modes on people's attitudes.