We conducted a few studies investigating the utility of explainable artificial intelligence (XAI)-assisted decision-making. This talk will present our efforts in using XAI to improve human-decision making. Rapid AI developments have enabled AI to support people's decisions in several domains, such as finance, law, medicine, and manufacturing. As such, there has been a remarkable increase in research on XAI-assisted decision-making and explainable AI's role in improving human-AI team performance. Studies are reporting both positive and negative effects of explanations. In our studies, we find that XAI has limited impact on improving users' trust, while there is some potential for explanations to improve users' task performance. We will discuss the potential reasons for our findings and conclude with an agenda for future research to improve XAI-assisted decision-making.

Speaker bio

Ronal Singh is a Research Fellow in Human-Agent Collaboration in the School of Computing and Information Systems. Ronal completed his PhD in 2018 from the University of Melbourne and his BSc and MSc degrees in Computer Science from the University of the South Pacific in Fiji Islands. Ronal's research interests include multi-modal human-agent interactions, explainable AI, intention recognition, and multiagent communication planning.