Communicating the Limitations of AI: The Effect of Message Framing and Ownership on Trust in Artificial Intelligence

Trust plays an essential role in the interaction between humans and artificial intelligence (AI). While too much trust could harm the interaction between humans and AI, the lack of trust in AI could increase users’ reluctance to use the system. One way to foster trust in AI is to provide information about the performance. However, the manner in which the information is presented, such as message framing and the ownership of decisions, could also affect the trust. To test this, we conducted a 2 (ownership: no ownership vs. ownership) × 3 (message framing: no information vs. negative information vs. positive information) between-subjects experiment. Participants were required to choose items for survival in the desert supported by AI decision aid. The results demonstrated that participants trusted AI more willingly in the absence of any information about the performance and when they did not have ownership of their decisions. Implications are discussed.