Why sterile processing stays with people
The question behind this page is simple: will AI replace medical equipment preparers, or just change parts of the shift? The honest answer is that the job is mostly hands. Instruments arrive soiled from the operating room. Someone has to sort them, brush lumens, run them through a washer-disinfector, then inspect each piece for cracks, dull edges and leftover bioburden under a lamp. Software can read a count sheet. It cannot feel a loose hinge on a clamp.
The second half of the day is assembly and delivery. Trays are built to a count sheet, wrapped or containered, loaded into a steam sterilizer or a low-temperature unit, then stored and carried to the right room at the right time. Each of those steps has a physical object in it. That is why the task mix on this page leaves so little room for software to work alone, and why hospitals keep staffing the department around people rather than machines.
The labor market has not thinned out either. The US Bureau of Labor Statistics counts about 77,420 medical equipment preparers, with median pay near $47,700 a year (BLS, 2025), and projects employment growth of roughly 10.8% between 2025 and 2035. Surgical volume drives the work. More procedures mean more trays, and trays still need turning around between cases. For how this role sits beside other hospital support jobs, see the hospitals sector page.
What software handles, what it assists, and what stays manual
The share AI can handle on its own is 0% of task time. That slice is paperwork: logging sterilizer cycles and load contents, and keeping inventory and supply records straight. Instrument tracking systems already scan trays, stamp the cycle data and flag a recall when a biological indicator fails, with no one typing it out.
Assisted work accounts for 7%. Here a tool speeds a person up without taking the task. Tray assembly is the clearest case: scanning instruments against a digital count sheet catches a missing retractor faster than a memorized list. Scheduling case carts and restocking par levels is the other, where demand forecasting tells the department what tomorrow’s board needs. The total of what any tool can do today is the coverage figure, 4 out of 100; the method behind it is explained on the coverage scoring page.
Everything else, 93% of task time, is the job as most people picture it. Decontaminating and hand-cleaning soiled instruments sits there. So does inspecting equipment for defects and pulling it from service, loading and unloading sterilizers, and setting up or delivering equipment in a procedure room. Our robotics review rates the physical side as fixed automation: washers and autoclaves run a set cycle in a fixed place, and a person still loads them, unloads them and decides what passes.
What the evidence says
There is no direct head-to-head test of an AI system against a trained sterile processing technician in this job. That is why the parity evidence grade is D, our lowest grade, and why no parity number appears on this page. Grading a job without a measured result would be a guess, and we do not publish guesses. The scoring rules are set out in the methodology.
What would settle it is specific. A study timing instrument inspection by a vision system against a technician, measured on missed defects and missed bioburden. A trial of automated tray assembly in a working central sterile department, scored on count-sheet accuracy and turnaround per tray. Published error and recall rates before and after an instrument tracking rollout at named hospitals. Until something like that exists, the fair reading is that the paperwork has been automated and the handling has not.
Good to know: machine sterilization is decades old, so the cycle running itself is not new automation; the open question is whether anything can take over sorting, inspection and assembly.
When the picture could change
Most likely after 2046 (8 in 10 of our scenarios). What that window measures is explained on the replacement-year method page.
Two things could pull it earlier. General-purpose robot arms that can grip varied, wet, oddly shaped instruments would attack the biggest manual block at once. And standardized, machine-readable instrument sets across manufacturers would make automated identification and tray building far easier than it is with mixed legacy inventory.
Two things hold it back. The capital cost of robotic handling in a department that already runs on inexpensive labor and fixed machines is hard to justify, and the hourly cost comparison on this page shows the gap. Then there is accountability: infection control failures are traced to a named person and a documented process, and hospitals have little appetite for shifting that responsibility to a machine that inspects a clamp.
How to stay needed in this role
Lean into the parts of the job that stay manual. Inspection and defect judgment is the first: knowing when an instrument is dull, pitted or out of alignment is a skill, not a scan. Decontamination for complex devices is the second, especially flexible scopes and lumened instruments with a long manual cleaning protocol. The third is being useful at the point of care, delivering and setting up equipment in a room under time pressure.
Two skills raise your floor. Certification and current knowledge of sterilization standards makes you the person who signs off a load rather than the person who loads it. Fluency with the instrument tracking system makes you the one who audits the data, investigates a failed indicator and trains new hires on it.
What to do: ask to own one high-complexity service line, such as endoscopy or orthopedics, and learn its trays end to end.
Nearby paths are worth comparing before you move. Endoscopy technicians work the same scopes from the procedure side. Medical equipment repairers go deeper into the devices themselves. Medical assistants trade equipment work for patient contact. You can put any two side by side with the job comparison tool, see the wider group on the other healthcare support occupations page, or check where hands-on roles land on the list of jobs least exposed to AI.