Trusted AI and the Contribution of Trust Modeling in Multiagent Systems

Researchers in the field of artificial intelligence today are increasingly concerned with whether the systems which they build will be "trusted AI", in other words, whether they will be accepted by their human users. The claim of this paper is that these researchers should be aware of the rich set of solutions being developed in the multiagent systems subfield of trust modeling. We propose a specific perspective on how to leverage trust modeling solutions towards assurances for trusted artificial intelligence. We conclude by advocating for greater dialogue between these AI communities.