EXPOSED
The control-room half of this job — reading gauges, trending process variables, adjusting setpoints, logging batch records — has been eroding for decades via DCS and advanced process control, and AI-assisted optimization accelerates that. The plant-floor half survives: lining up valves, connecting hoses and rail cars, pulling samples, clearing plugged lines, and physically walking a unit through startup and upset conditions in a hazardous environment where robots aren't deployed. No licensure protects the role, so the pressure shows up as fewer operators covering more units rather than the job disappearing.
Headcount grew steadily across the period.
Median pay $49,130 → $58,040 -5.5% 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
+3.3% 128,900 → 133,100 on the projections basis
Exposed, but growing
AI can already do a lot of these tasks, and the BLS still expects +3.3% more of these jobs by 2034. Demand for the output is growing faster than the work is being automated away — the mechanism BLS gives for software developers, and the combination people most often misread as an error.
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.
~14,400 openings a year on average, including replacing people who leave.
SublimerVaporizerAcid MakerSalt MakerSoda MakerAcid TenderCausticiserCell TenderChlorinatorLye TreaterNeutralizerPond TenderResin MakerAcid CleanerAcid RemoverAcid TreaterPlant WorkerSalt RefinerAcid BleacherAcid OperatorAcid PurifierCell OperatorPoly OperatorSoda Dialyzer
Holding it up: embodiment . Weakest point: trust premium .
Mixed — a routine tier and a judgment tier Setpoint trending, batch-sheet logging and reactor charging sequences are already recipe-automated on any modern DCS, but manual valve line-ups, hose and rail-car hookups, grab sampling, filter changes and unplugging a fouled transfer line keep roughly half the shift off the console — hence 11 rather than the 6 of a pure board-watching job or the 15 of a millwright.
Hands-on in uncontrolled environments You are in Nomex, gloves and a respirator on catwalks and tank tops in whatever weather, breaking flanges on lines that held caustic or solvent, and no vendor sells a robot that will walk a unit through a startup or chase a leak at 2 a.m. — the 15 rather than 18 reflects that a real share of hours are spent seated at a screen in a climate-controlled control room.
Certification preferred, not legally required OSHA PSM 1910.119 and RCRA obligations land on the facility and the process safety engineer, not on you; your credentials are typically a HAZWOPER card, confined-space and lockout/tagout training and an employer-run qualification checklist, all of which the next hire can complete in weeks — nothing that a regulator would have to revoke.
Meaningful discretion You call an emergency shutdown, decide whether a rising reactor temperature is instrument error or a runaway, and choose the sequence to depressurize — real discretion with real consequences, but bounded by written SOPs, the batch recipe and a supervisor or engineer reachable by radio, which is what puts this at 9 instead of the mid-teens.
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 (5/20) is whether the law requires a licensed human to sign. Trust premium (4/20) is whether buyers specifically pay for a person. Judgment and accountability (9/20) 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 18 of this occupation's 44 points (41%).
Embodiment (15/20) 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 57/100, still EXPOSED.
Task-mix shift plus formal recognition: as APC and AI optimizers absorb steady-state setpoint management, the residual role concentrates on upset response, abnormal-situation management, and go/no-go calls on continuing versus emergency-shutdown. If site procedures and union contracts name the console operator as the person with unilateral shutdown authority (the 'stop work authority' language already in many refinery and USW contracts, strengthened post-Chevron Richmond), the accountability is documented rather than implicit
Genuine two-tier structure: routine trending, logging, and batch-record entry automate away, leaving field verification, sampling under nonstandard conditions, plug-clearing, and walking a unit through startup — tasks AI cannot perform and which resist proceduralization because each unit's quirks are undocumented. Score rises by subtraction, not by new capability barriers
OSHA Process Safety Management (29 CFR 1910.119) and EPA Risk Management Program revisions — the RMP reconsideration rulemaking is live — could require a named, qualified human operator to authorize and sign each safe-work permit, line break, and startup/shutdown step, with explicit prohibition on automated systems executing PSM-covered operating-procedure deviations without human concurrence. CSB recommendations after AI-assisted-control incidents are the likely trigger. Also plausible: state boiler/pressure-vessel or ammonia-refrigeration operator certification requirements (already exist for stationary engineers in some jurisdictions) extended to chemical process units
Little headroom — already 15. Could edge up if insurers or PSM rules restrict mobile robot deployment in Class I Div 1 electrically classified areas, keeping field rounds human; conversely explosion-proof inspection robots (Ex-rated ANYbotics, Taurob already in refinery trials) push this down
The limit. No route to trust_premium: the buyer is a corporate plant owner minimizing headcount per unit, and no customer of the chemical product knows or pays for who tended the reactor. Even with a strong liability shield, the realistic ceiling is roughly 55-60 — the shield would protect the existence of a signing operator per unit, not the number of operators, so the displacement pressure continues to appear as span-of-control expansion rather than job elimination.
| Houston-Pasadena-The Woodlands, TX | 9,390 | $79,430 +37% |
| New York-Newark-Jersey City, NY-NJ | 8,060 | $54,790 -6% |
| Chicago-Naperville-Elgin, IL-IN | 4,510 | $61,250 +6% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 4,110 | $60,410 +4% |
| Dallas-Fort Worth-Arlington, TX | 3,640 | $47,230 -19% |
| Baton Rouge, LA | 2,500 | $93,920 +62% |
| Indianapolis-Carmel-Greenwood, IN | 2,330 | $61,020 +5% |
| San Juan-Bayamon-Caguas, PR | 2,300 | $36,450 -37% |
| Beaumont-Port Arthur, TX | 1,540 | $96,790 +67% |
| New Orleans-Metairie, LA | 1,330 | $94,450 +63% |
| Baton Rouge, LA | 2,500 | $93,920 +62% |
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 44. 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.