EXPOSED
Modal worker here runs a fossil, hydro, or combined-cycle plant from a control room: monitoring SCADA displays, adjusting load and steam parameters, logging readings, and coordinating with dispatch — all of which DCS automation and predictive analytics have been eating for two decades, long before LLMs. What holds is the physical half: field rounds, manual valve and breaker switching, lockout/tagout, startup and shutdown sequences, and diagnosing why a pump is cavitating when instrumentation disagrees with reality. The bigger threat to this occupation isn't AI at all — it's plant retirements and the shift to unstaffed or remotely-operated generation, which is already shrinking headcount per megawatt.
This fall is concentrated in 2020 and has not recovered since.
Median pay $81,990 → $102,040 -0.4% 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
-11.2% 31,600 → 28,000 on the projections basis
Shrinking, but not obviously because of AI
The BLS projects -11.2% by 2034, but at 50/100 this work is only moderately exposed — not the profile of a job current AI can simply do. Occupations shrink for many reasons, and the score does not point at automation as this one's cause.
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.
~2,500 openings a year on average, including replacing people who leave.
OperatorRectifierFuel HandlerUnit OperatorEnergy ManagerHydro MechanicHydro OperatorPlant MechanicPlant OperatorPower OperatorShift OperatorStation TenderTurbo OperatorRelief OperatorBooster OperatorChiller OperatorControl OperatorPlant TechnicianStation OperatorTurbine OperatorElectric OperatorPowerhouse TenderAuxiliary OperatorChiller Technician
Holding it up: embodiment . Weakest point: trust premium .
Mixed — a routine tier and a judgment tier Load-following, setpoint adjustment, and log-keeping are already handled by the DCS with the operator supervising, but the trip-and-recovery work — cold starts, condenser vacuum problems, chasing a bad transmitter against physical evidence — still needs someone in the room, which puts it mid-band rather than in single digits.
Hands-on in uncontrolled environments Field rounds through turbine halls and boiler decks at 100°F+, hanging LOTO tags, racking breakers, stroking valves by hand when air is lost, and reading local gauges that aren't wired to anything are unavoidable and happen in a plant that is noisy, hot, and pressurized — well past a hybrid control-room job.
Certification preferred, not legally required Most fossil and hydro operators hold NERC system-operator certification or a state boiler/turbine operator license and are named on the shift log, but the plant manager and the utility carry the regulatory exposure, so the credential constrains who can hold the seat without making the individual the personal defendant.
Meaningful discretion Deciding to trip a unit versus riding through an abnormal vibration or low drum level is a call worth millions in equipment and it lands on the operator in seconds, but most of those decisions are bounded by written alarm-response procedures and operating limits that specify the action.
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 (8/20) is whether the law requires a licensed human to sign. Trust premium (5/20) is whether buyers specifically pay for a person. Judgment and accountability (12/20) 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 25 of this occupation's 50 points (50%).
Embodiment (14/20) 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.
Ship Engineers SAFE
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 68/100 — SAFE.
Genuine two-tier structure: as DCS and predictive analytics absorb steady-state load-following and logging, the residual job concentrates in abnormal-condition diagnosis, instrument-versus-reality conflicts, and non-routine startup/shutdown. This raises task resistance per remaining worker without any new law — but note it operates alongside headcount contraction, so the surviving job is harder to automate while there are fewer of them.
NERC extending the operator-certification model (currently NERC System Operator certification is required for transmission/balancing-authority desks, not generation control rooms) to require a certified, named individual to authorize generation dispatch changes and startup/shutdown sequences — plus an explicit rule that AI advisory output cannot be the authorizing signature. A narrower version already exists in 10 CFR 50.54(m) for nuclear licensed operators, which is why nuclear operators score far higher; a fossil/hydro analogue would be the specific thing to watch.
State PUC or insurer conditions on remote/unstaffed operation permits requiring a licensed operator physically on site during startup, black-start, and switching evolutions — the same structure FM Global and boiler-inspection codes (ASME CSD-1, state boiler-attendant laws in e.g. Michigan and Pennsylvania) already impose on attended high-pressure boilers. Erosion of those attendant laws is the live downside risk; their extension to combined-cycle and BESS-paired plants is the upside.
Grid-reliability rules that name the plant operator as the accountable party for ride-through, frequency-response, and black-start performance during declared emergencies, with post-event NERC compliance investigation attaching to the individual's log entries — the pattern that followed Winter Storm Uri FERC/NERC inquiries. If enforcement attaches to named operators rather than to the plant entity, the role owns consequential calls in a way automation cannot absorb.
Interconnection of large battery, hydrogen co-firing, or synchronous-condenser retrofits adds unfamiliar field hardware with immature instrumentation, increasing manual isolation, LOTO, and hands-on commissioning work per plant during the retrofit window.
The limit. No plausible route to trust_premium: electricity is a commodity and no buyer can identify, let alone pay extra for, a human in the control room. More importantly, every lever above raises resistance per remaining worker while the dominant force — plant retirement and MW-per-operator consolidation toward remote operations centers — shrinks the count. A higher score here does not mean more jobs.
| New York-Newark-Jersey City, NY-NJ | 1,090 | $131,770 +29% |
| Boston-Cambridge-Newton, MA-NH | 640 | $103,900 +2% |
| Chicago-Naperville-Elgin, IL-IN | 640 | $123,540 +21% |
| Los Angeles-Long Beach-Anaheim, CA | 540 | $105,870 +4% |
| Houston-Pasadena-The Woodlands, TX | 530 | $104,620 +3% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 470 | $99,000 -3% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 380 | $62,080 -39% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 310 | $101,350 -1% |
| Wenatchee-East Wenatchee, WA | 70 | $145,510 +43% |
| Fresno, CA | 130 | $138,930 +36% |
| Boise City, ID | 70 | $134,050 +31% |
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 50. 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.