Interaction Mediation for Meaningful Human Control over Highly Automated Vehicles

Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics, pp. 4411–4416

Abstract

The current advancement of technology with not only physical, but increasing cognitive functions widely coined as Artificial Intelligence (AI) leads to multiple situations where humans willingly or unwillingly accept the decisions of algorithms and Neural Network (NN) Models resulting in humans being decreasingly involved in the decision-making process. Human-machine cooperation and Human Autonomy Teaming (HAT) are an answer to this problem, joining the best of both sides into joined cognitive systems. Meaningful Human Control (MHC) is the concept of ensuring that humans have enough influence on the outcome of actions in HAT and hence stay morally responsible and legally accountable for their actions. This paper extends the concept of using an interaction mediator that facilitates the interaction between an automated and a human agent with interaction patterns to maintain MHC. The concept is applicable in almost any domain, and will be shown with the example of guiding a highly automated vehicle.

Authors 3

1. Marcel Caspar Attila Baltzer

   Fraunhofer Institute for Communication, Information Processing and Ergonomics

   Affiliation as printed

   Fraunhofer FKIE,Wachtberg,Germany

   Fraunhofer FKIE, Wachtberg, Germany

2. Alexander Ripkens

   Fraunhofer Institute for Communication, Information Processing and Ergonomics

   Affiliation as printed

   Fraunhofer FKIE,Wachtberg,Germany

   Fraunhofer FKIE, Wachtberg, Germany

3. Frank Ole Flemisch Aachen Dept. balanced HSI

   RWTH Aachen University · Fraunhofer Institute for Communication, Information Processing and Ergonomics

   Affiliation as printed

   Fraunhofer FKIE,Wachtberg,Germany

   Dept. balanced HSI, RWTH Aachen University, Aachen, Germany

   Fraunhofer FKIE, Wachtberg, Germany