Efforts should target clear approval processes, research-driven assurance methods, and workforce development. Education and policy codification are essential for scaling ML application approvals, with system-level strategies offering alternatives to item-level complexities by validating behavior regardless of implementation.
Low-risk and some higher-risk ML applications already have approval pathways, forming a defensible baseline. Education, outreach, and operational experience within this baseline are crucial to refining methods and raising the assurance standard. Civil aviation’s strong safety record reflects decades of prioritizing safety, and any new methods must align with regulatory safety intent, proving their value before adoption.
For the full version of the paper published in and presented at the 43rd Digital Avionics Systems Conference, San Diego, September 2024 click here.



