Caution with AI deception: AI could serve as the scapegoat of Human!

In the era of widespread artificial intelligence (AI), it is crucial to understand the distinct roles of AI and humans in deception. This study employs an innovative empirical approach to investigate how deceived responses extend from human deception to artificial intelligence deception. We conducted an experiment using virtual scenario manipulations to explore how individuals perceive and respond to deception when humans and AI act as solo or collusive deceptive agents. The results indicate that when both humans and AI engage in deception under false propositions, AI is perceived as more deceptive and intentional than humans, making it more likely to bear the blame and receive reduced trust. The willingness to communicate with AI also experiences a decline. The perception of intent to deceive serves as a mediating factor in the connection between the agent (human or AI) and the attribution of blame or trust, rather than the presence of actual deception. This study provides empirical evidence for understanding AI deception and highlights the ethical and moral concerns when humans collaborate with AI and depend on AI-generated information.