Towards certifiable AI in aviation: landscape, challenges, and opportunities

TL;DR

This paper explores the landscape, challenges, and opportunities for certifying AI systems in aviation, emphasizing the need for qualification beyond just performance metrics to ensure safety and robustness.

Contribution

It provides a comprehensive mind map of formal AI certification in avionics and highlights key challenges and considerations for certifying AI in safety-critical aviation systems.

Findings

Certification challenges are significant for AI in avionics.

Qualification requires more than performance metrics.

A need for new standards and approaches in AI certification.

Abstract

Artificial Intelligence (AI) methods are powerful tools for various domains, including critical fields such as avionics, where certification is required to achieve and maintain an acceptable level of safety. General solutions for safety-critical systems must address three main questions: Is it suitable? What drives the system's decisions? Is it robust to errors/attacks? This is more complex in AI than in traditional methods. In this context, this paper presents a comprehensive mind map of formal AI certification in avionics. It highlights the challenges of certifying AI development with an example to emphasize the need for qualification beyond performance metrics.