The Intersection of Robust Intelligence and Trust: Hybrid Teams, Firms and Systems

Abstract

We are developing the physics of interdependent uncertainty relations to efficiently and effectively control interdependence for autonomous hybrid teams (i.e., arbitrary combinations of humans, robots and machines), which cannot be done presently. Uncertainty is created in states of interdependence between social objects: at one extreme, interdependence reduces to independent agents and certainty but with asocial, low-power solutions generating little meaning or understanding in social contexts; oppositely, the length of interdependence increases across a group, de-individuating its members until individual identity dissolves (e.g., cults, gangs, well-run teams), increasing power, efficiency and meaning internal to a group, but also the chances of mal-adaptation (e.g., tragic mistakes). We focus on how interdependence increases the robust intelligence of a group by increasing its autonomy while decreasing its entropy, but requiring external control to be indirect. For humans, teamwork is an unsolved theoretical problem; solving it should generalize to the effective computational control of hybrid teams, a path forward for the users of a team to trust it to operate safely in hostile environments. Present theories of interdependence, like game theory or social science, are inadequate to formulate strategies to control teams; alternative theories like machine learning can control swarms with pattern formations, but not states of interdependence, such as multi-tasking operations. While alternative theories cannot be used to model teams, decision-making or social conflict at the same time (hostile mergers; checks and balances), ours can.