Challenges and Future Directions for AI-based Systems for Remote Digital Towers

Remote Digital Towers (RDTs) are reshaping Air Traffic Management (ATM) by enabling controllers to manage operations through high-definition video, audio, and sensor data from remote facilities. While offering operational flexibility and efficiency, RDTs also raise new challenges around workload, reliability, and trust—especially as Artificial Intelligence (AI) becomes increasingly integrated into decision-support systems. This study, conducted within the SESAR JU–funded TRUSTY project, explores how AI, including multimodal machine learning, generative and explainable AI (XAI), can augment rather than automate RDT operations. We present a state-of-the-art review of AI in ATM, highlight lessons learned from prior projects (e.g., ARTIMATION, MAHALO, etc.), and propose a multimodal solution combining video, audio, and rulebased methods for event detection and decision support. A validation exercise with air traffic controllers tested different levels of explainability and transparency, showing how adaptive interfaces can influence trust, workload, and decision-making. Results indicate that AI can enhance situational awareness and resilience but must be designed with transparency, controllability, and calibrated trust to avoid overreliance or misuse. Building on these findings, the paper provides recommendations for advancing AI- enabled RDTs, including multimodal data fusion, adaptive explainability, human–AI teaming, and cross-project SESAR alignment. The study concludes that AI in RDTs should act as a second pair of eyes, extending vigilance and supporting ecological and operational goals, while keeping human controllers firmly in command.