SAFE
The core of this job is pulling low-voltage cable through finished walls, ceilings, and attics, mounting devices, terminating panels, and troubleshooting nuisance alarms in buildings that never match the as-built drawings — none of which current robotics can do. AI does compress the paperwork tier: device schedules, battery/voltage-drop calculations, permit packages, inspection reports, and customer quotes are increasingly generated automatically. Fire alarm work carries a real regulatory moat (state low-voltage licensing, NICET certification, AHJ acceptance testing signed by a qualified person), which is thinner on the residential security-install side where the modal worker may be doing DIY-adjacent camera and sensor setups.
Dipped in 2020, then grew past where it started.
Median pay $48,970 → $60,070 -1.9% 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
+10.4% 85,900 → 94,900 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +10.4% 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.
~9,400 openings a year on average, including replacing people who leave.
InstallerAlarm AdjusterAlarm InstallerAlarm TechnicianSecurity EngineerInstall TechnicianSecurity InstallerService TechnicianFire Alarm MechanicFire Alarm InstallerElectronic TechnicianSurveillance OperatorBurglar Alarm MechanicBurglar Alarm InspectorBurglar Alarm InstallerInstallation TechnicianSurveillance TechnicianSecurity Alarm InstallerAccess Control TechnicianHome Fire Alarm InstallerSecurity System InstallerControl Systems TechnicianProtective Signal RepairerProtective Signal Installer
Holding it up: embodiment . Weakest point: trust premium .
Tasks largely resist digitisation Fishing 18/2 through a plaster wall with no chase, drilling fire-rated top plates, and chasing an intermittent ground fault on a 60-device initiating circuit are all improvised, site-specific work; the 15 rather than 18 reflects that the design-side tasks — device layout from a floor plan, NFPA 72 spacing math, voltage-drop and standby battery sizing, panel programming templates — are now largely software-generated, so the drafting and takeoff portion of your week is genuinely compressible.
Hands-on in uncontrolled environments You spend the day in attics above 120°F, on ladders and lifts at ceiling height, crawling above suspended tile in occupied buildings, and pulling wire in active construction with other trades in the way — an 18 rather than 20 only because a slice of the work is bench-level panel programming and head-end/NVR configuration you can do sitting down.
Licensed human required and personally liable Fire alarm acceptance and annual inspection/test records under NFPA 72 have to be signed by a qualified person, and most states gate the work behind a low-voltage or fire alarm license with NICET Level II expected on commercial jobs — the 12 not higher because the license usually attaches to the company's contractor registration rather than your personal seal, and residential burglar-alarm and camera installs often need only a registration card or nothing at all.
Meaningful discretion Deciding whether a repeatedly tripping smoke is dust, a bad detector, or a real fire condition — and whether to leave a system in trouble, place it on fire watch, or notify the AHJ — is a call you make on site with life-safety consequences and no script; the 13 rather than 15+ because NFPA 72, the approved shop drawings, and the manufacturer's listing constrain most of your choices about placement, wire type, and circuit configuration.
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 (15/20 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 (12/20) is whether the law requires a licensed human to sign. Trust premium (10/20) is whether buyers specifically pay for a person. Judgment and accountability (13/20) 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 35 of this occupation's 68 points (51%).
Embodiment (18/20) 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 78/100, still SAFE.
Task-mix shift is genuine here: as quoting, device schedules, permit packages and inspection reporting get generated automatically, the residual day is retrofit cable routing, nuisance-alarm root-cause troubleshooting, and integration of mismatched legacy panels — the tier that resists. Also raised if mass sensor/IoT retrofit and NFPA 72 remote-inspection rules increase the volume of commissioning and false-alarm diagnostic work per building.
States that currently exempt low-voltage security/alarm work from licensing adopting NICET Level II or equivalent as a condition of permit signoff, mirroring the fire-alarm side — e.g. a state fire marshal rule requiring the acceptance-test certificate (NFPA 72 Record of Completion) to be signed by a named certified technician personally, not just the contracting firm. Also plausible: insurers writing residential policies conditioning burglary/fire credits on installation by a licensed technician rather than self-install.
Codified responsibility for design decisions in retrofit: if AHJs push back on AI/software-generated device layouts and voltage-drop calcs by requiring a human technician to attest that field conditions match the submitted design (a 'field verification' line on the permit package), the installer owns the ambiguity call rather than executing a plan.
Only a narrow route: commercial and institutional buyers (hospitals, schools, data centers) specifying named certified technicians in bid documents. Residential DIY-adjacent segment has no realistic path — Ring/SimpliSafe-style self-install pressure runs the other way.
The limit. Embodiment at 18 is already near the practical ceiling; retrofit cable pulling in unmapped buildings is among the hardest remaining physical tasks, so there is nothing to add there. The real downside risk is not AI but product design — cellular, battery, self-commissioning devices shrinking install labor per site.
| New York-Newark-Jersey City, NY-NJ | 6,850 | $59,280 -1% |
| Dallas-Fort Worth-Arlington, TX | 3,550 | $57,930 -4% |
| Los Angeles-Long Beach-Anaheim, CA | 2,960 | $71,570 +19% |
| Houston-Pasadena-The Woodlands, TX | 2,500 | $56,920 -5% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 2,220 | $60,560 +1% |
| Phoenix-Mesa-Chandler, AZ | 1,730 | $60,750 +1% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 1,610 | $66,790 +11% |
| Chicago-Naperville-Elgin, IL-IN | 1,510 | $61,310 +2% |
| San Jose-Sunnyvale-Santa Clara, CA | 320 | $81,980 +36% |
| Salem, OR | 70 | $81,500 +36% |
| Kahului-Wailuku, HI | 30 | $81,340 +35% |
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.