SAFE
A police sergeant or lieutenant spends the shift responding to scenes, taking incident command, making use-of-force and arrest-authority calls, and handling personnel discipline — none of which a model can own. The automatable slice is real but administrative: report review and correction, shift scheduling, crime-stat compilation, and drafting routine memos and grant paperwork. Sworn status and statutory authority mean a certified human must sign the arrest approval, the use-of-force review, and the pursuit termination decision.
Headcount grew steadily across the period.
Median pay $91,090 → $106,040 -6.9% 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
+2.9% 160,800 → 165,400 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +2.9% 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.
~10,900 openings a year on average, including replacing people who leave.
CaptainLieutenantPort WardenChief DeputyPolice ChiefDesk SergeantDeputy SheriffPatrol CaptainPolice CaptainPost CommanderDetective ChiefPatrol SergeantPolice SergeantPatrol CommanderPrecinct CaptainShift SupervisorTraffic SergeantPolice LieutenantPolice SupervisorSafety SupervisorDetective SergeantTraffic LieutenantTraffic SupervisorHarbor Patrol Chief
Holding it up: judgment & accountability . Weakest point: trust premium .
Tasks largely resist digitisation Approving a felony arrest at 0300, taking incident command on a barricaded-subject call, and sitting a subordinate down for a Garrity-warned interview are the bulk of the shift, and the software-friendly residue — CAD report kickbacks, roll-call scheduling, monthly UCR/NIBRS compilation — is why this sits at 14 and not 18.
Hands-on in uncontrolled environments Supervisors roll to the scene rather than watch it: foot pursuits, crowd lines, MVA and homicide scene control, physically supervising cell blocks and evidence handling, all in weather and traffic and hostile crowds, which puts it at 16 rather than 20 only because a real share of hours are spent at a desk on reports and CompStat prep.
Licensed human required and personally liable POST/state certification is the statutory precondition for the sworn authority you exercise, your name goes on the use-of-force review and the pursuit-termination order, and §1983 supervisory-liability and failure-to-train claims land on you personally — decertification, not just termination, is the exposure.
Exists to be accountable for ambiguous calls Deadly-force authorization, terminating a pursuit into a school zone, deciding whether a use-of-force incident goes to IA or ends with counseling, and a Graham v. Connor reasonableness call made in seconds with no supervisor above you on scene at night — these are unreviewable in the moment and litigated for years after.
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 (14/20 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 (16/20) is whether the law requires a licensed human to sign. Trust premium (12/20) is whether buyers specifically pay for a person. Judgment and accountability (18/20) 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 46 of this occupation's 76 points (61%).
Embodiment (16/20) 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 83/100, still SAFE.
State POST (Peace Officer Standards and Training) boards or consent-decree language explicitly requiring a named supervisor of record to review and personally certify every AI-assisted report, license-plate-reader hit, or facial-recognition identification before it supports probable cause — already the pattern in Detroit's 2024 ACLU settlement over facial recognition, which mandates supervisory corroboration and bars FR alone as an arrest basis. Broad adoption across states would raise this.
Post-Floyd duty-to-intervene and supervisory-liability statutes (e.g., Colorado SB20-217, Illinois SAFE-T Act) plus DOJ consent decrees naming sergeants as accountable for subordinate force reviews; further erosion of qualified immunity for supervisors, or state civil-rights acts allowing direct suit against supervisors, would tighten this.
Task-mix shift: if Axon Draft One-style report generation and automated scheduling/CompStat tools absorb the administrative tier, the residual shift is scene command, force review, and discipline — none currently modelable. Countervailing: report review is itself a core sergeant task, so automation of drafting increases review load rather than removing it.
The limit. Near the practical ceiling at 76. Trust premium has no route: policing is a monopoly public service with no buyer choosing between a human and machine supervisor, so willingness-to-pay for humanness cannot be expressed. Realistic gains are confined to liability formalization and the judgment tier.
| New York-Newark-Jersey City, NY-NJ | 13,970 | $135,270 +28% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 5,050 | $125,960 +19% |
| Los Angeles-Long Beach-Anaheim, CA | 4,840 | $167,600 +58% |
| Atlanta-Sandy Springs-Roswell, GA | 4,460 | $81,860 -23% |
| Chicago-Naperville-Elgin, IL-IN | 3,800 | $135,080 +27% |
| Boston-Cambridge-Newton, MA-NH | 3,720 | $108,190 +2% |
| Baltimore-Columbia-Towson, MD | 3,440 | $124,660 +18% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 3,250 | $119,540 +13% |
| San Francisco-Oakland-Fremont, CA | 1,690 | $175,840 +66% |
| Vallejo, CA | 150 | $175,070 +65% |
| Santa Cruz-Watsonville, CA | 70 | $172,590 +63% |
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 76. 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.