EXPOSED
The core work — setting roll gaps, threading stock, monitoring gauges, adjusting speed and tension, measuring thickness with micrometers and calipers — is hands-on inside a hot, loud mill, which language AI cannot touch and current robots handle poorly. The real threat isn't ChatGPT, it's closed-loop mill automation: automatic gauge control, thickness sensors, and PLC recipe libraries that already absorb the tending and fine-adjustment tasks and let one operator cover several lines. Production reporting, scrap logging, and spec lookup are the first things to disappear; the physical setup, changeover, and diagnosing why the strip is crowning or wandering are what keep people on the floor.
Most of this decline happened after 2021 — it is not the pandemic dip.
Median pay $40,490 → $50,140 -0.9% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
-8.3% 22,500 → 20,600 on the projections basis
Exposed, and shrinking
Both signals point the same way: the tasks are largely automatable and the BLS projects -8.3% by 2034. This is the case where the score and the forecast agree, and it is the one worth taking seriously.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~1,900 openings a year on average, including replacing people who leave.
RollerSetterFlangerLevelerManglerRougherScarferFinisherBar RollerBulldoggerHot RollerCold RollerManipulatorRoll SetterTube RollerBed OperatorBrass RollerGuide SetterRip OperatorRod FinisherSteel RollerStrip RollerTable RunnerBarrel Roller
Holding it up: embodiment . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier Automatic gauge control and stored PLC recipes already handle steady-state tending and trim adjustments, but roll changeouts, threading a new coil, shimming and crowning corrections, and reading why the edge is wavy still need a person with a wrench — that split is what puts it at 11 rather than in the resistant band.
Hands-on in uncontrolled environments You work inside the mill: climbing the stand to change work rolls, feeding stock through pinch rolls, wiping coolant and scale, hand-miking strip at temperature — uncontrolled enough for 13, but on a fixed line with known geometry rather than a shifting outdoor site, which is why it isn't 17.
No licence, no signature requirement No state licence gates roll operation; a plant may want a forklift or overhead crane card and OSHA lockout/tagout training, but the mill owns product liability and no certificate follows you as personal legal exposure.
Meaningful discretion You decide when a roll is worn enough to pull, whether off-gauge strip gets scrapped or downgraded, and how much tension to give a thin gauge before it snaps — real calls made against a written spec and a supervisor's sign-off, not open-ended ones, hence 8.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (11/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (2/20) is whether the law requires a licensed human to sign. Trust premium (3/20) is whether buyers specifically pay for a person. Judgment and accountability (8/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 13 of this occupation's 37 points (35%).
Embodiment (13/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 48/100, still EXPOSED.
Closed-loop gauge control absorbs steady-state tending, leaving the job as changeover setup, roll/die swaps, and diagnosing shape defects (crowning, camber, strip wander) that AGC cannot correct — a genuine two-tier split where the routine tier is what automates. Watch for job postings that drop 'tender' language and require roll-shop/setup and defect-diagnosis skill.
Customer/qualification regimes rather than licensure: AS9100 or NADCAP aerospace flow-down, or ASME/API pressure-vessel material qualification, requiring a named qualified operator to sign the mill certification and process sheet for each heat/coil. Already partially in force for aerospace and nuclear-grade material; extension to more product lines would formalize a signed human record.
OSHA machine-guarding and lockout/tagout enforcement plus insurer requirements make a named human responsible for authorizing entry into the roll bite for jam clearing and for the decision to run vs. scrap a coil — the accountability sits with the operator even when the control loop is automated. Would rise further if a facility's automated line runs with only one authorizing human per shift.
Shift toward high-mix, short-run specialty alloy and thin-gauge plastic film work, where every changeover is a physical re-thread and roll re-crown in a hot mill; robotic threading of coil ends remains unsolved. Reshoring of specialty rolling under Section 232 tariffs would push mix this way rather than toward long automated runs.
The limit. No realistic route to a trust premium — rolled coil and film are sold on certified dimensional and metallurgical spec, and no buyer pays extra to know a human set the roll gap. Also note that liability and judgment gains here concentrate value in a shrinking headcount: one accountable setter covering several automated lines raises that role's resistance while eliminating the tenders around it, so occupation-level employment can fall even as the score rises.
| Chicago-Naperville-Elgin, IL-IN | 1,300 | $49,850 -1% |
| Pittsburgh, PA | 1,260 | $69,150 +38% |
| Atlanta-Sandy Springs-Roswell, GA | 730 | $47,730 -5% |
| Reading, PA | 670 | $70,310 +40% |
| Los Angeles-Long Beach-Anaheim, CA | 570 | $48,080 -4% |
| Louisville/Jefferson County, KY-IN | 550 | $50,400 +1% |
| Detroit-Warren-Dearborn, MI | 500 | $59,660 +19% |
| Dallas-Fort Worth-Arlington, TX | 490 | $57,280 +14% |
| Decatur, AL | 70 | $114,470 +128% |
| Reading, PA | 670 | $70,310 +40% |
| Pittsburgh, PA | 1,260 | $69,150 +38% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 37. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.