LO
Air traffic controller — how we know
The page itself gives the judgements. This one gives what they rest on: which technologies bear on the work, how the estimate moved since language models reached the public, and the method behind both.
Which technologies matter here#
Four separate signals. They are deliberately not added together — a job exposed to two technologies is not twice as exposed.
How it got here#
The index is not a static number. This is where it would have sat at each capability checkpoint since ChatGPT — reconstructed, and labelled as such.
—— this stretch contains a verified event- - - no event in this stretch — reconstruction only0 = no task exposed, 100 = every task exposed
● 1 verified event for this occupation, plotted at the date it happened — the parts of the line near a marker are anchored to something checkable.
The flattest curve on this site, and the shape is the finding rather than a lack of data. It starts at a level that has nothing to do with language models: conflict alerting, radar data processing and electronic flight data have been machine work for decades, and the number would have looked similar in 2015. Almost nothing happens across the checkpoints where other occupations move steeply, because the capability class that changed — reading, drafting, answering — is not the class this job is gated on, and the gate is a licence issued to a named person for a named position, which changes by rulemaking rather than by release note. The small late rise is two specific things that are documented rather than assumed: simulators taking a share of training that previously had to happen on live traffic, and the agency's own stated intention to point machine learning at flow and demand ahead of the day of departure. Both are about the work around the position; neither is about the decision made at it. Read the line's height, not its slope: a low number that barely moves is a different statement from a high number that has stopped rising.
A flat line is not a forecast of safety. It says which tasks automation has reached so far — the occupations that moved least here are the ones where the constraint is physical or regulatory, and both of those can change.
Method and sources#
- Assessment date
- 2026-09-23
- Basis of the task judgements
- 2 evidence-backed · 4 platform inference · 0 not enough evidence
- Verified events
- 2