User Participation in the Design of Trustworthy Human-AI-Collaboration in Air-Traffic Control

Kurzfassung

The introduction of AI-based systems is one of the core elements in the course of digital transformation in aviation. Human factors expertise is needed to find user-centered approaches of human-AI collaboration. The DLR project Collaboration of human operators and AI systems (LOKI) aims at developing concepts, demonstrators and prototypes for trustworthy human-AI collaboration in air traffic control. To consider users’ perspective in early design phases, two workshops with experienced air traffic controllers of Deutsche Flugsicherung GmbH (DFS) and Austro Control were conducted. The metaplan method was used to asses users’ expectations and their requirements on Human-AI collaboration. In both workshops, ten air traffic controllers participated. The workshops highlighted what users expect from trustworthy Human-AI collaboration in Air Traffic Control. The main results were users’ assessment which of their tasks could be delegated to an AI system and how the delegation of tasks to AI-systems should be designed. The workshops were a success among the participants as they provided insight into future requirements, responsibilities and tasks. Furthermore, they helped outline potential challenges in the interaction concept that must be solved to integrate AI systems successfully in aviation. These findings will be used for the concept design of prototypes.