EXPOSED
The core of the job — being physically present, walking a site, checking doors, and responding when something is wrong — is not automatable with today's robotics, and that keeps guards employed. But the single largest slice of guard hours is camera monitoring and access logging, and AI video analytics already flags intruders, loitering, weapons, and left-behind objects better than a human on hour six of a night shift. Expect the occupation to persist in headcount but shift away from monitor-watching posts toward mobile patrol, incident response, and customer-facing lobby roles, with each remaining guard covering more square footage.
Dipped in 2020, then grew past where it started.
Median pay $29,680 → $38,020 +2.5% 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
+0.4% 1,262,100 → 1,267,100 on the projections basis
Growing, and only partly exposed
The BLS expects +0.4% more of these jobs by 2034, and at 53/100 the work is only partly exposed — some tasks are automatable, the core of the job is not. Nothing here is in tension.
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.
~161,000 openings a year on average, including replacing people who leave.
BouncerGatemanWatchmanBodyguardPatrolmanBank GuardCamp GuardDoorshakerGate GuardWatchguardArmed GuardBag CheckerDoor TenderFloorpersonFloorwalkerGate KeeperGate PersonGate TenderHall TenderPlant GuardShip KeeperTruck GuardDriver GuardFlex Officer
Holding it up: embodiment . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier A 12 reflects the split within one job description: foot patrols of stairwells, unlocking doors for a locked-out tenant, escorting a hostile ex-employee off the floor, and standing a fixed post at a hospital ED all still need a body, but the console shift — scanning 40 camera feeds, logging badge-ins, writing the shift report from a template — is exactly what video analytics and automated access control already do, and that's often half or more of the paid hours at a large site.
Hands-on in uncontrolled environments A 17 is because the work happens in parking garages at 2 a.m., on construction sites in rain, in warehouse yards with uneven ground and trucks moving, and in crowds where you may physically block, restrain, or wrestle with someone who does not want to be moved — nothing about that is a controlled environment, and the only reason it isn't higher is that a meaningful share of guards spend their shift seated at a lobby desk.
Certification preferred, not legally required A 7 fits the reality that most states require a guard card or registration (California BSIS, New York's 8-hour pre-assignment, Florida Class D) with fingerprinting and a background check, which is a real gate an algorithm can't pass — but it's a weeks-long licence, not a professional one, and when a guard uses excessive force or misses an incident the negligent-security suit lands on the contract firm and the property owner, not on the guard's personal licence.
Meaningful discretion An 8 is set by post orders: guards operate on written observe-and-report instructions with explicit escalation to police or a supervisor, and most decisions are procedural — yet the calls that matter (is this person a threat or just agitated, do I intervene in a fight or hold the door, do I evacuate now) are made alone in seconds with no one to consult, which is why this sits above the procedural band rather than in it.
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 (12/20 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 (7/20) is whether the law requires a licensed human to sign. Trust premium (9/20) 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 24 of this occupation's 53 points (45%).
Embodiment (17/20) 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.
Flight Attendants 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 70/100 — SAFE.
Use-of-force and detention decisions concentrating in fewer remaining posts as monitoring is automated: the guard who executes a citizen's arrest, escorts an involuntary discharge, or de-escalates an armed subject owns a call no vendor will indemnify. Healthcare and school-security contracts increasingly specify certified de-escalation (CPI, IAHSS) as a condition of the post.
Insurer requirements: property and casualty carriers conditioning premises-liability coverage (or negligent-security defense) on documented licensed-guard presence and human sign-off on incident reports, as some carriers already do for armed posts and hospital/school contracts.
Genuine two-tier structure: if camera-watching hours are absorbed by analytics, the residual role is patrol, response, and ambiguous-encounter handling, which is not the same job measured against current benchmarks. Watch for post orders that eliminate the console seat and consolidate hours into mobile response.
State licensing boards (e.g., California BSIS, New York DOS Division of Licensing) requiring that any AI-generated alarm verification or trespass determination used as the basis for a police dispatch be confirmed and logged by a licensed guard — mirroring existing state 'alarm verification' statutes that already require human confirmation before police response to reduce false alarms. Extension of those verification mandates from alarm companies to video analytics would put a licensed name on each dispatch.
Buyer segments that explicitly purchase visible human presence: luxury retail, houses of worship after threat incidents, hospital EDs, and union-negotiated staffing floors (SEIU Local 32BJ contracts specify minimum staffed posts in NYC commercial buildings). Expansion of such contract floors, or municipal ordinances requiring staffed lobbies in large residential buildings, would raise this.
The limit. Most of these protect a smaller number of posts rather than aggregate headcount; the console-monitoring tier has no lever and its erosion is not reversible by any of the above. Trust premium is real only in specific high-consequence or high-luxury segments, not in the low-bid contract-guard market that employs the majority.
| New York-Newark-Jersey City, NY-NJ | 142,760 | $43,980 +16% |
| Los Angeles-Long Beach-Anaheim, CA | 83,280 | $40,030 +5% |
| Chicago-Naperville-Elgin, IL-IN | 48,360 | $39,990 +5% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 41,530 | $36,040 -5% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 40,950 | $46,830 +23% |
| Dallas-Fort Worth-Arlington, TX | 34,000 | $36,680 -4% |
| Houston-Pasadena-The Woodlands, TX | 29,250 | $35,460 -7% |
| San Francisco-Oakland-Fremont, CA | 26,830 | $45,730 +20% |
| Idaho Falls, ID | 570 | $78,680 +107% |
| Fairbanks-College, AK | 130 | $60,400 +59% |
| Kennewick-Richland, WA | 500 | $54,670 +44% |
The New York Post reports that a Boston Dynamics Spot robot is being used to patrol and monitor a construction site on Long Island, New York.
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.