■ HIGH RISK ■ Transportation & Logistics
The gangs are shrinking, not disappearing. Mechanized track machines and drone inspections keep cutting crew sizes, but somebody still has to fix a broken rail at 2 a.m. in the rain, and it won't be a robot anytime soon.
“Track-laying machines don't need spike-driving contests.”
Our AI replacement risk score — how we score jobs
Track workers — maintenance-of-way crews, in railroad speak — keep thousands of miles of rail safe to run trains on. The work ranges from tie replacement and ballast tamping to rail grinding, welding broken joints, aligning track geometry, and emergency repairs after derailments or washouts. It's heavy, outdoor, safety-critical labor governed by strict protocols, because the job site is an active railroad and the thing that kills track workers is usually a train.
Mechanization has been eating this job for seventy years, and the pace hasn't slowed. Modern tamping machines, ballast regulators, and tie-replacement gangs do with a dozen operators what once took a hundred laborers. Inspection is the newest front: geometry cars measure track alignment at line speed, drones survey bridges and rights-of-way, and machine-vision systems flag defective ties and fasteners from moving vehicles, replacing the walking inspector with a data feed. Railroads under perpetual pressure to cut operating ratios treat maintenance headcount as a target, and predictive-maintenance algorithms increasingly decide where the remaining crews go.
What survives is everything unpredictable. Broken rails in winter, storm damage, derailment cleanup, thermite welds in awkward locations, and work in terrain where the big machines can't reach — that's human work, and it's not automating on any near horizon because the environments are unstructured and the stakes are absolute. The job is also unionized and safety-regulated, which slows workforce cuts even when technology allows them. Expect fewer, more skilled positions: machine operators, welders, and signal-adjacent technicians rather than large section gangs. Our 62 score reflects steady attrition through technology rather than sudden replacement.
Automatability: our editorial assessment of current and near-term AI capability
The disruption is a long slope, not a cliff, and it's steepening now. Automated inspection is displacing walking inspectors this decade, and each generation of track machine needs a smaller gang. Serious workforce pressure lands by 2030 as predictive maintenance concentrates work into fewer, machine-heavy crews — but emergency and repair work keeps a human core employed well beyond that.
Secure-ish, shrinking slowly. Railroads will always need maintenance-of-way crews for repairs and emergencies, and unions plus federal safety rules slow the cuts. But crew sizes fall with every equipment generation, and automated inspection is thinning inspector ranks now. The safest positions are machine operators and welders, not general laborers.
Dramatically. Geometry cars and machine-vision systems mounted on trains now detect defects at line speed, and drones handle bridges and hard-to-reach territory. That turns inspection from a walking job into a data-review job — fewer boots on ballast, more defects found. The finding is automated; the fixing still isn't.
Collect qualifications. Machine-operator certifications, welding tickets, and signal or bridge cross-training all move you from the replaceable end of the gang to the durable end. Railroads face genuine shortages of skilled operators and welders even while trimming laborer positions, so the ladder exists — climb it before your class of job gets consolidated.
Not in any realistic planning horizon. Machines excel at planned, repetitive work on accessible track. They're helpless at a washout at midnight, a broken rail in a snowstorm, or a derailment site. Unstructured emergency work in bad conditions is precisely what robotics is worst at, and railroads run on exactly that kind of chaos.