Why this job stays close to the road
Traffic technicians sit between a desk and a work zone. Part of the job is data: tabulating vehicle counts, checking cycle lengths, pulling crash and volume records into a report for an engineer. Software is good at that part, and has been for a while. The other part happens at an intersection, in traffic, often with a cone truck parked behind you.
Someone has to place and retrieve the counters, tubes and portable cameras, then check that the loop or detector is actually reading cars. Someone has to stand at the corner during a signal timing change and watch what the queue does. When a resident calls about a left-turn arrow that is too short, or a school crossing that feels unsafe, the follow-up is a site visit and a conversation, not a prompt.
The scale matters too. The Bureau of Labor Statistics counted about 7,860 traffic technicians in the United States, with median pay near $59,090 a year, and projects roughly 4% employment growth from 2025 to 2035 (BLS). This is a small, mostly public-sector occupation. Cities add signals faster than they add staff, which tends to push technicians toward more intersections each, not fewer.
What software does, what it assists with, what stays with people
Start with the share of task time our scoring puts in each group. The tasks marked as work AI can handle are the paperwork end of the job: compiling counts and speed data into tables, and generating the routine summaries and permit or study documents that follow a field survey. That slice is 4% of task time. Adaptive signal systems also recalculate timing plans from detector data without a person touching each plan.
The assisted group is larger in practice than most people expect. Video analytics can classify and count vehicles from a camera feed, but a technician still sets the camera, checks the classification against reality, and signs off. Drafting signing and striping plans is similar: the layout tools do the geometry, the technician decides what the intersection needs. Assisted work accounts for 43% of task time.
Then the field. Installing and retrieving counting equipment, inspecting signal heads and detectors, coordinating with electricians and street crews, and meeting the public about a complaint all sit in the group that needs a person: 53% of task time. Overall coverage, the share of task time AI can handle today, reads 25 out of 100. If you want the definition behind that figure, see how coverage is scored.
What the evidence actually shows
Here is the honest position. No published study has tested an AI system against a working traffic technician on this job’s full task set. The quality parity grade is D, which is our marker for not measured, so no parity number is given on this page and none should be guessed from adjacent research.
Plenty of transportation research measures pieces of the work, such as signal optimization algorithms or camera-based counting accuracy against manual counts. That is useful, but it tests a tool on one task under controlled conditions. It does not test whether a system can run a corridor retiming from complaint to field verification without a technician. What would settle it is a field trial: same corridor, same study requests, software-led versus technician-led, scored on accuracy, rework and public complaints over a season. Until something like that is published, the grade stays where it is. Our full approach is set out in the scoring methodology.
When the balance could shift
Most likely after 2043 (8 in 10 of our scenarios). For what that window measures and how it is built, read the replacement year method.
Two things could pull it earlier. First, connected-vehicle and probe data reduce the need to put hardware in the pavement at all, which removes field trips rather than automating them. Second, state and city agencies that already run centralized signal operations centers can push more intersections per technician, which thins entry-level roles before it touches senior ones.
Two things hold it back. The physical share of this job needs hardware at the dexterous humanoid tier, as the robotics panel above shows, and that tier is not deployable on a live roadway today. And public agencies carry liability: a signal change that contributes to a crash becomes a legal and political problem, so a named person signs the plan. Procurement cycles in city and state DOTs are slow by design, which stretches adoption out further.
Good to know: the pressure here shows up first as fewer junior count-and-tabulate positions, not as fewer intersections needing attention.
How to stay needed
Lean into the work the task list above leaves with people. Field verification of detectors and timing changes is the obvious one, because it is where software’s output meets a real queue. Public and interagency contact is the second: complaint investigations, school-zone reviews, and coordinating a retiming with police and transit. Third, inspection and acceptance of contractor signal and signing work, where judgment and liability travel together.
Two skills raise the floor. One is intelligent transportation systems work, including detector and controller diagnostics, communications networks and central software administration. The other is data review: knowing when a camera count or an adaptive timing plan is wrong, and being able to show why. The current Still needs a human score, 74 out of 100 (higher is safer), rests heavily on that field half of the job.
Nearby roles worth looking at include Transportation Inspectors, Civil Engineering Technologists and Technicians and Electrical and Electronic Engineering Technologists and Technicians. You can also browse the wider other transportation workers family, see where the transportation and warehousing sector sits, put two roles side by side with the job comparison tool, or check the list of jobs that mostly need a person.