Accountability, Authority, And The Data Infrastructure That Makes Ethical AI Possible In Wildland Firefighting.
Most people working in fire operations have formed a view on AI. Whether that view is broadly positive, cautious, or somewhere in between, it was almost certainly shaped by experiences that have little to do with aerial firefighting: a consumer tool that impressed, a headline that alarmed, a demonstration that felt either compelling or unconvincing. That is a reasonable way to form an initial impression of a new technology. It is a less reasonable basis for procurement decisions, governance frameworks, and accountability structures in a life-safety environment.
This paper is about the gap between those two things. What the science tells us about how humans actually perform alongside automated systems. What operational data from real fire seasons reveals about the reliability of the inputs feeding those systems. And what the industries that have already navigated this challenge learned, mostly at cost, about what accountable AI deployment actually requires.
