Operational explainability of AI: Development of a crew interaction system for civil aircraft​

Kurzfassung

This poster presents an overview of the work on operational explainability of AI at DLR, Institute of Flight Guidance. Operational explainability is the ability of a technical system to explain AI-based results in a way that allows human end users, such as pilots, to understand how the AI calculated the system's results while performing flight‑deck operational tasks. This work described aims to explore, develop, implement and validate a concept for AI‑operational explainability for airline flight‑deck systems. The following research questions were addressed: How should the AI Crew Interaction System (AICIS) be designed as an AI‑based decision‑support system in the cockpit in order to ensure operational AI explainability?