Why this work stays at the bedside
Will AI replace licensed vocational nurses? The task mix answers most of it. An LPN or LVN spends the day measuring and recording vital signs, administering prescribed medications, starting IV fluids, changing dressings and helping patients bathe, dress and move. That is physical work on a changing human body, done under a state license, in a room where the plan often changes mid-shift.
Software can read a chart. It cannot feel a limb that is warm and swollen, notice that a usually chatty resident has gone quiet, or hold a confused patient steady while a catheter is placed. Most of the measured task time here is physical, and the robotics tier shown above is a dexterous humanoid: hands that work in tight, wet, unpredictable spaces. That hardware is not in nursing homes at scale, and the cost comparison on this page shows why swapping care staff for machines is not a cheap trade.
There is also accountability. Medication administration and sample collection are delegated acts inside a legal scope of practice. A tool can flag a dose that looks wrong. Someone licensed still has to decide, give it and sign for it. You can see how that shapes the headline figure on the Still needs a human score page.
What AI does, what it helps with, and what it leaves to nurses
A small slice of the work already runs without a nurse in the loop (share of task time: 0%). It is clerical: recording food and fluid intake and output into the record, and pushing routine notes and vitals into the chart from connected monitors. Nothing in that group touches a patient.
A larger group is assistance (share of task time: 4%). Ambient documentation tools draft shift notes from speech. Barcode and dose-check systems cross-read medication orders before administration. Scheduling and handoff tools sort who needs what next. The nurse still does the act; the software shortens the typing and catches some errors.
The rest stays with people (share of task time: 96%). Wound and ostomy care, repositioning and basic hygiene, assembling and using equipment such as catheters and oxygen supplies, and talking a family through what is happening all sit here. That is why overall coverage, our answer to whether AI can do the job today, lands where it does: 8 out of 100. The coverage method page explains how task time is weighted.
What the evidence actually tests
Not much, in this job. Our evidence grade for quality parity is D. Published work has compared AI systems with clinicians on reading text, drafting notes and answering exam-style questions. None of it measures an AI system doing an LPN’s shift: giving a scheduled injection, spotting early skin breakdown, calming an agitated resident at 3 a.m.
So there is no parity number here, and we do not publish one. What would settle it is a trial in a real facility that tracks the same outcomes nursing already tracks: medication error rates, pressure injury rates, falls, documentation accuracy and time to escalate a change in condition, with and without the tool, graded by an independent reviewer. Until that exists, the honest read is a documentation story, not a care story. How grades are assigned is set out on the quality parity page, part of our open method.
Good to know: a D grade means nobody has measured this job against people yet, not that AI scored poorly.
When the picture could shift
Most likely after 2042 (8 in 10 of our scenarios). The replacement-year method explains what that window measures and how the scenarios are built.
Two things could pull it earlier. First, cheap tooling: documentation and monitoring software costs a fraction of a staffed shift, so facilities adopt it fast, and each adoption moves more charting out of the job. Second, remote monitoring that handles vitals and early-warning alerts could cut the number of nurses needed per unit without replacing any single nurse.
Two things hold it back. Hands are the first: the physical share of this job needs humanoid dexterity in cluttered, fluid-filled, unpredictable settings, which is covered in our guide to humanoid robots and physical jobs. Licensing and liability are the second: scope-of-practice rules name a licensed person as the one who administers and signs. Demand matters too. BLS counts about 648,410 LPN and LVN jobs in the United States, median pay of $64,400, and roughly 3% employment growth projected for 2025 to 2035 (BLS). Aging care needs are growing while facilities already struggle to staff shifts.
How to stay needed as an LPN or LVN
Lean into the work that sits in the human group. Three worth building on: complex wound and ostomy care; physical assessment of patients whose condition is changing, where noticing comes before measuring; and the family and resident conversations that keep a care plan workable. Those are the tasks facilities protect when budgets tighten.
Two skills pay off alongside them. Learn to supervise the documentation tools rather than fight them, so you can catch a wrong draft note or a bad dose flag fast. And get comfortable with the devices: pumps, monitors, remote-monitoring dashboards and the alerts they throw.
Registered nurses is the usual next step, and an LPN-to-RN bridge program is the common route. Psychiatric technicians sits close in task terms if behavioral health interests you, and nurse practitioners is the longer path. You can put any two of them side by side on our compare tool, see the wider health technologists and technicians family, or check how the rest of nursing homes and residential care scores. For a broader view of hands-on roles, our list of jobs that mostly need a person is a good place to start.