An initial Framework for the development of Artificial Intelligence (AI) in Aviation

Abstract

The aviation sector is increasingly exploring the application of Artificial Intelligence (AI) technologies to enhance operational efficiency, safety, and overall system reliability. This thesis presents a novel framework for the development, testing, and implementation of AI within the aviation industry, underpinned by comprehensive guidelines that accommodate technical, regulatory, ethical, economic, and human factors. Central to this research is the integration of the Intelligent Pilot Advisory System (IPAS), a Level 1b AI-based system developed by the German Aerospace Centre (DLR), aimed at supporting pilots in high-workload scenarios. The framework proposes stringent safety measures, robust technical validations, and compliance with established aviation standards to ensure that AI systems enhance rather than compromise operational integrity. Empirical data from a simulator study involving twelve certified pilots highlights the practical benefits of IPAS, demonstrating significant improvements in decisionmaking efficiency during abnormal flight situations. Moreover, the study revealed a positive shift in pilot perception towards AI technologies, underscoring their potential to reduce workload and enhance decision accuracy significantly. However, the integration of AI in aviation is not devoid of challenges. The thesis discusses several barriers, including data security concerns, the complexity of system integration, and the crucial need for explainable AI outcomes to build user trust and understanding. Regulatory backgrounds are also in flux, necessitating a framework that can evolve alongside emerging AI technologies and changing policy environments. In conclusion, while AI offers transformative potential for aviation, its successful implementation will require a balanced approach that addresses both the opportunities and inherent challenges. The proposed framework serves as a foundational step in guiding stakeholders through the landscape of AI integration, ensuring that the benefits of AI are harnessed to enhance the efficiency and safety of aviation operations.