In AI We Trust: How Humans Teach an AI to Reciprocate Trust

Synopsis

We study whether humans trust an Artificial Intelligence (AI) when they are given the opportunity to teach the AI how to reciprocate the trust of a human being. The AI initially has four strategies to reciprocate trust: minimum reciprocity, slight reward for trusting, fair sharing of benefits, and generosity. Subjects can teach or unteach these strategies to the AI according to the outcomes in each period. Results indicate that subjects rapidly unteach the two least profitable strategies to the AI, then figure out the most profitable strategy.

Speaker

Dr Benjamin Prisse is a Research Fellow at the Lee Kuan Yew Centre for Innovative Cities (LKYCIC) at the Singapore University of Technology and Design (SUTD). He did a Master of Economics at Toulouse School Economics, then a Master of Neuroeconomics at Maastricht University. He recently completed a PhD in Data Science with a particular focus on Experimental Economics at University Loyola Andalucía. His research interests are visual experiments, online environments, Artificial Intelligence, experiments with teenagers.