A risk-based trust framework for assuring the humans in human-machine teaming

Trust in human-machine teaming (HMT) is a difficult concept to define and manage. This is largely due to its subjectivity. Despite its subjectivity, trust has become a critical element of safety for HMT. The level of trust a human has in an autonomous or AI-enabled system impacts the decisions they make when operating alongside that system. As HMT involves a greater interdependence between humans and autonomous or AI-enabled systems, there is a greater significance on the human roles and responsibilities in this dynamic. Humans are an equal, if not greater contributor to trustworthy HMT. This paper presents a human centric risk-based trust framework for HMT. The framework is intended to be used to support risk assessments for HMT which captures human centric trust related risks and enforces measures for facilitating calibrated levels of trust in HMT operations. The presented framework focusses on interventions for the human in HMT specific to the implementation phase.