Field Note / 2026-05-25
AI in GIS without the smoke machine
A draft field note on where AI may help geospatial work: cleanup, classification, routing, summaries, QA, and decision support without pretending judgment can be outsourced.
This note is a parking lot for the GIS essay track: where AI is already useful, where it is still hype, and where geospatial professionals still need to own the judgment.
The useful questions are practical: Can the tool clean messy data? Can it help classify imagery? Can it summarize field notes? Can it catch bad attributes? Can it support a map product without hallucinating authority?
The standard stays the same: if a model touches a map, the human still owns the decision, the metadata, and the proof.