SAFE
The job is climbing into energized substations, bucket trucks, and generating stations to test protective relays, calibrate breakers, trace faults in switchgear, and commission equipment under live-grid constraints — none of which robotics can do at cost today. AI will absorb the paperwork tier: relay setting calculations, test-report generation, SCADA alarm triage, and trend analysis of transformer and battery data. What it cannot absorb is diagnosing an intermittent trip on a 138kV line at 2am and deciding whether to re-energize, a call with blackout and arc-flash consequences attached to a named human.
Most of this decline happened after 2021 — it is not the pandemic dip.
Median pay $82,780 → $103,020 -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
+5.5% 23,400 → 24,700 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +5.5% more of these jobs by 2034. Note that safe does not mean well paid — several of the fastest-growing resistant occupations are among the lowest paid on the register.
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.
~2,000 openings a year on average, including replacing people who leave.
LinemanWiremanInside WiremanRelay EngineerRelay OperatorGear TechnicianMeter CraftsmanRelay CraftsmanTest TechnicianMeter TechnicianRelay TechnicianApparatus LinemanGenerator MechanicSubstation WiremanPowerhouse MechanicSubstation MechanicCorrosion TechnicianTransformer RepairerElectrical TechnicianMotor Room ControllerSubstation TechnicianInstrumentation FitterMaintenance TechnicianPowerhouse Electrician
Holding it up: embodiment . Weakest point: trust premium .
Tasks largely resist digitisation Injecting test current into a differential relay with an Omicron set, torque-checking breaker linkages, and pulling CT secondaries to prove polarity are physical verification steps that cannot be simulated away — the automatable slice is the setting math and the test sheet, not the wrench and the leads, which is why this sits at 16 rather than in the mixed band.
Hands-on in uncontrolled environments You work inside energized 138kV yards, on top of transformers, in cramped relay houses and generating-station cable vaults, in ice, heat, and at 2am callout — a 19 reflects that the environment is not just physical but genuinely uncontrolled, with arc-flash boundaries and switching orders governing where your body can be.
Certification preferred, not legally required There is no state licence for relay techs the way there is for a PE or a journeyman electrician in some jurisdictions; you hold NERC-relevant qualification records, employer switching authorization, and often NETA/NICET certification, so the utility and the stamping engineer carry the formal liability while your name still lands on the test report — that combination is a 10, not a 15.
Exists to be accountable for ambiguous calls Deciding a misoperation was a CT wiring error versus a genuine fault, whether to leave a bus on a single protection scheme overnight, or to re-energize after an intermittent trip are calls made with incomplete oscillography and blackout consequences, but they sit inside switching orders, NERC PRC standards, and an on-call engineer you can escalate to — hence 15 rather than 18.
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 (16/20 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 (10/20) is whether the law requires a licensed human to sign. Trust premium (8/20) is whether buyers specifically pay for a person. Judgment and accountability (15/20) 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 33 of this occupation's 68 points (49%).
Embodiment (19/20) 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 82/100, still SAFE.
State PUC or OSHA-adopted NFPA 70E rule requiring a documented, individually-signed energized-work permit and arc-flash risk assessment by a qualified person for every live substation task, closing the current practice of blanket utility permits; some investor-owned utilities' own safety manuals already do this internally
Task-mix shift: once AI absorbs relay-setting calculation, test-report generation and SCADA alarm triage, the residual job is almost entirely commissioning novel equipment (inverter-based resources, HVDC converter stations, grid-scale storage) where settings interact unpredictably and no training corpus exists — this occupation genuinely has a paperwork tier and a live-diagnosis tier
NERC expanding PRC-005 / PRC-027 protection-system compliance to require a named, qualified technician (or licensed PE) to attest to relay settings-verification records, with personal attestation retained in the audit trail rather than utility-level sign-off — the direction FERC/NERC audit findings on protection misoperations already push toward
Post-event enforcement practice (as in FERC/NERC misoperation penalties and the after-action reviews following Texas 2021 and Moss Landing-type events) formally attributing re-energization decisions to the named on-site technician of record rather than to a control-center dispatcher, which raises the ambiguity-ownership of the role
Utility mutual-aid and insurance requirements specifying human-verified commissioning for interconnection of new generation — narrow, since the buyer is a utility procurement department, not a retail customer who can express a preference for humans
The limit. Embodiment at 19 and task_resistance at 16 are near ceiling; realistic headroom is 8-11 points, almost all in liability_shield, which is currently low only because accountability sits with the utility entity rather than the individual. The trust_premium dimension has essentially no route above ~10: the customer is a monopoly utility, not a discretionary buyer, so no willingness-to-pay-for-human mechanism exists. Note the countervailing risk: fleet-wide substation automation and remote-testing retrofits reduce headcount without touching any of these scores.
| New York-Newark-Jersey City, NY-NJ | 2,870 | $96,170 -7% |
| Los Angeles-Long Beach-Anaheim, CA | 540 | $118,990 +16% |
| Houston-Pasadena-The Woodlands, TX | 520 | $103,940 +1% |
| Dallas-Fort Worth-Arlington, TX | 410 | $102,470 -1% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 400 | $106,510 +3% |
| Austin-Round Rock-San Marcos, TX | 390 | $103,120 +0% |
| Boston-Cambridge-Newton, MA-NH | 290 | $119,460 +16% |
| Charlotte-Concord-Gastonia, NC-SC | 280 | $105,300 +2% |
| Fresno, CA | 40 | $145,530 +41% |
| San Francisco-Oakland-Fremont, CA | 90 | $141,580 +37% |
| Seattle-Tacoma-Bellevue, WA | 220 | $135,130 +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 68. 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.