SAFE
The bulk of an RN's shift is physical and interpersonal: starting IVs, assessing wounds, repositioning patients, administering meds, catching the subtle change in color or breathing that precedes a crash. AI is already eating the documentation layer — charting, discharge summaries, care-plan drafts, triage protocols, insurance paperwork — which is real time savings but not the job. State licensure, personal accountability for medication errors, and the fact that patients and families want a human at the bedside keep this occupation firmly intact; the near-term risk is not replacement but higher patient ratios justified by 'AI efficiency'.
Headcount grew steadily across the period.
Median pay $73,300 → $97,550 +6.5% 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
+4.9% 3,391,000 → 3,557,100 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +4.9% 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.
~189,100 openings a year on average, including replacing people who leave.
NurseWard NurseField NurseScrub NurseStaff NurseX-Ray NurseCharge NurseCounty NurseFlight NurseSchool NurseTrauma NurseTriage NurseClinical NurseDelivery NurseDialysis NurseForensic NurseGenetics NurseNeonatal NurseOncology NursePrenatal NurseSurgical NurseVascular NurseVisiting NurseAdmission Nurse
Holding it up: embodiment . Weakest point: judgment & accountability .
Tasks largely resist digitisation At 14 the score credits that hanging a piggyback, titrating pressors to a MAP target, doing a neuro check q1h, and de-escalating a confused post-op patient at 3am cannot be done through a screen — but it stops short of 18 because charting, care-plan generation, SBAR handoff drafts, discharge instructions, and acuity/triage scoring are genuinely automatable and can eat 20-30% of a shift.
Hands-on in uncontrolled environments 19 reflects that the work happens in rooms with vomit, code carts, combative patients, and bariatric transfers — you cannot start a 20-gauge in a dehydrated 88-year-old, palpate a rigid abdomen, or feel a thready radial pulse remotely, and the physical unpredictability of a med-surg floor is why this sits at the ceiling rather than in the 13-15 range of clinic-based roles.
Licensed human required and personally liable State board licensure with a personal NPI, a named signature on every MAR entry, and the reality that a wrong-patient insulin dose goes to the Board of Nursing under your license — not the hospital's — puts this at 16, held below 19 only because RNs practice under physician orders and protocols rather than holding independent prescriptive authority.
Exists to be accountable for ambiguous calls Deciding whether the drop in urine output warrants waking the intensivist, whether to hold a beta blocker at a borderline pressure, or whether this patient is septic before the lactate returns are ambiguous calls with defined consequences — 14 rather than 18 because much of that discretion runs through standing orders, rapid-response criteria, and escalation to a physician who owns the final diagnostic decision.
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 (17/20) is whether buyers specifically pay for a person. Judgment and accountability (14/20) 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 47 of this occupation's 80 points (59%).
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 92/100, still SAFE.
State safe-staffing laws with hard numeric RN-to-patient ratios that explicitly bar counting AI monitoring or documentation tools toward staffing compliance — California Title 22 already sets ratios; Oregon HB 2697 (2023) and pending bills in NY/MI/PA extend the model. A ratio statute that names AI acuity-scoring as non-substitutable would block the 'AI efficiency = higher ratios' pathway that is the actual near-term risk.
Task-mix shift: this occupation genuinely has two tiers. As charting, discharge summaries, care-plan drafts and insurance documentation are absorbed by ambient scribes (Epic/Nuance DAX deployments already live at Kaiser, HCA), the residual shift concentrates in escalation decisions, rapid-response calls, titration under ambiguous vitals, and family goals-of-care conversations — the parts that own consequences. No law required.
Expansion of formal RN authority under protocol — nurse-driven sepsis bundles, RN-initiated Foley removal, standing-order titration, and independent triage disposition authority in EDs. Where hospitals codify these, the RN owns a consequential call previously made by a physician.
Board of Nursing rules (or NCSBN model language) declaring that an RN who accepts an AI early-warning score, sepsis alert, or AI-drafted assessment without independent verification is practicing below standard of care — making the nurse the required verifying signer on algorithmic output rather than a downstream consumer of it. Malpractice insurers writing this into policy conditions has the same effect.
If bedside AI absorbs the routine tier faster than staffing shrinks — i.e., ratio laws hold — the remaining hour-by-hour work is unstructured physical assessment and de-escalation, which current systems cannot do at usable quality. Rises only conditional on the staffing floor; without it, saved documentation time is reclaimed as more patients, not more judgment.
Growth of segments where the human presence IS the product and is separately billed: hospice/palliative bedside care, private-duty and concierge nursing, doula-adjacent perinatal nursing, and home-based hospital-at-home programs (CMS Acute Hospital Care at Home waiver) that require an in-person RN visit per day by rule.
The limit. Already 80/100 with embodiment at 19 — very little headroom. The realistic fight is defensive: preventing task_resistance erosion via staffing floors, not raising the score. A single-payer-style cost squeeze or repeal/preemption of ratio laws moves this down faster than any lever moves it up.
| New York-Newark-Jersey City, NY-NJ | 197,740 | $119,720 +23% |
| Los Angeles-Long Beach-Anaheim, CA | 109,360 | $135,560 +39% |
| Chicago-Naperville-Elgin, IL-IN | 100,240 | $100,490 +3% |
| Dallas-Fort Worth-Arlington, TX | 76,680 | $101,420 +4% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 73,790 | $101,180 +4% |
| Houston-Pasadena-The Woodlands, TX | 65,910 | $99,830 +2% |
| Boston-Cambridge-Newton, MA-NH | 64,240 | $106,180 +9% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 61,670 | $91,380 -6% |
| San Jose-Sunnyvale-Santa Clara, CA | 22,930 | $216,740 +122% |
| Vallejo, CA | 4,120 | $203,290 +108% |
| San Francisco-Oakland-Fremont, CA | 41,750 | $186,610 +91% |
Carle Health · NHS · Houston Methodist · WVU Medicine · Mercy · Minneapolis VA Healthcare System · Sharp HealthCare; MaineHealth · West Cumberland Hospital, Whitehaven · Montefiore Medical Center · Munson Healthcare
Houston Methodist's chief medical information officer described the health system's adoption and implementation of AI scribe technology for clinicians, in an interview with HealthLeaders Media.
The Guardian reports that nurses at New York hospitals have been displaced by AI systems, with nurses raising concerns about care quality.
Chief Healthcare Executive reports on WVU Medicine's experience deploying ambient AI documentation tools for clinicians.
A nurses' union says Montefiore plans to replace 12 nursing positions with AI following a strike, as reported by Crain's New York Business.
Spectrum News reports nurses at Munson Healthcare ratified a three-year contract that includes guardrails on the use of AI along with annual raises.
Healthcare IT News reports that Mercy health system has deployed ambient AI documentation tools for nursing workflows.
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.