A Shared Intelligence Framework Targeting Flexible Human Agent Cooperative Tasks

High quality decisions in military and commercial operations often require the combined intelligence of multiple agents, whose knowledge, expertise, and capabilities are complementary. Cooperative decision making and problem solving is hard for people who have different languages, priorities, and incentives, and is even more difficult with the participation of non-human agents with inscrutable behaviors to people. A fundamental challenge is in developing artificial agents that collaborate in a natural way with humans in order to utilize individual expertise to benefit the whole. Our innovation is a shared intelligence framework, a constrained, but still widely-applicable, role of human agent collaboration (HAC), for managing human-agent teams and ultimately facilitating both human-agent cooperative tasks and human-to-human activity coordination. Our approach addresses the overall problem of cooperative intelligent behavior by splitting it into two sections, where collaborative decision-making falls on the shoulders of human experts who work together (possibly through agent intermediaries), but where the generation and flexible adaptation of the expert team is conducted through HAC.