Decision making in human-machine teaming

This project aims to address the problem of accountable decision making in human-machine teaming scenarios. In human-machine teams there is a concern that humans could adopt a position of 'social loafing' (leaving the machine to do all the work). Alternatively they might do all the work, taking over from the machines at every possibility due to lack of appropriate trust. In an ideal situation, human-machine teams would lead to social facilitation, whereby the performance of both humans and machines is optimized. In this instance, human operators will only take control from machines when it is beneficial to do so. This project will explore accountable human decision making in human-human teams to provide an analogy for human-machine teaming. The template of accountable decision making will be used to derive principles for the design of human-machine teams and will include considerations such as task type, trust, workload, and situation awareness. An experimental environment will be constructed to enable exploration of accountable decision making, in which the factors such as number of machine agents, task allocation and workload can be manipulated to discover the effects on decision making. It is anticipated that the experimental programme will involve around six experiments to find a 'sweet spot' for accountable decision making. This 'sweet spot' will be turned into a set of design principles and guidelines for Thales.