FAA / Research report

DOT/FAA/TC-26/2

Safety Assurance for AI/ML in Safety-Critical Airborne Systems

Research

Edition: June 2026

Source checked
November 16, 2026

Open official publication record ↗

Aerospace Consulting / Reference note

Research into assurance approaches for airborne AI/ML, including verification, uncertainty, runtime assurance and contingency management. Useful for identifying technical questions around protection and recovery architectures.

A research publication, not an adopted certification method. A monitor or fallback still requires a substantiated role, coverage, timing, independence where credited, and appropriate assurance.

This page contains source metadata and selected explanatory notes. Use the official publication for full content, attachments and current status.

Discuss this reference for your programme →