Abstract

The present research examined the psychological mechanism that lead individuals to expect ethical responsibility of artificial intelligence (AI). Specifically, this research tested the hypothesis that the positive effect of anthropomorphism on the AI’s ethical responsibility is mediated by the perceived freewill. The findings showed that the higher the level of anthropomorphism of AI, the greater the AI’s ethical responsibility. Also, this effect was fully mediated by the perceived AI’s freewill caused by anthropomorphism. Additionally, findings showed that the positive effect of perceived freewill on ethical responsibility applies to both human and AI. That is, participants reported lower ethical responsibility for AI than human as the level of freewill of AI was perceived to be lower than that of human. This research offers theoretical implications of expanding the understanding of human-AI interactions by revealing that the psychological mechanism in which individuals expect ethical responsibility to AI is the perceived free will of AI. It also provides managerial and policy implications by suggesting the need for ethical design of AI for human-AI interaction.