Why the fitting room keeps a person in it
Ask will AI replace orthotists, and the answer sits in what the job actually involves. An orthosis or prosthesis is not a product pulled off a shelf. The clinician measures the residual limb or the joint, takes a cast or a digital impression, and then reads how the tissue changes under load. That reading happens with hands on a living body, in a room with a patient who is tired, sore, or anxious about walking again.
Two tasks carry most of the weight. The first is fitting and adjusting the device on the person: heat-shaping a socket, grinding a trim line, watching gait, then going back and changing it again. The second is instructing patients on how to wear, clean, and care for the device, which is part teaching and part negotiation. Software can propose a shape. It cannot feel a pressure point forming or decide how much the patient will tolerate this week.
The paperwork side is different. Writing up patient records, preparing documentation for payers, and building digital design files are all tasks where tools already do real work. That is the honest shape of change here: some tasks erode, the clinical core does not. Pay and demand give the job room to absorb that. The US median wage was $81,110 and employment about 9,390 (BLS, 2025), with projected growth of 12.9% from 2025 to 2035 (BLS, 2025).
What AI does, what it assists with, and what stays with people
The slice AI can take on its own is the documentation and digital-file work: dictated clinic notes turned into records, and scan data cleaned up into a usable model. That share of task time is 0%. It is the part of the day that happens at a screen rather than at a bench or a patient.
A second slice is assisted rather than handed over. Design work on a socket or a brace, and the ordering and tracking of materials and components, both go faster with software suggesting shapes and flagging errors, while the clinician signs off. That share is 20%. Here the tool shortens the loop; it does not close it.
Everything else sits with the clinician: evaluating the patient and taking the cast or measurement, fitting and adjusting the finished device, repairing and modifying it after wear, and training the patient to use it. That share is 80%. Because so much of the work is physical and patient-facing, the overall Can AI do it figure stays small; the way we count task time is set out in how coverage is measured.
What the evidence shows, and what is missing
No one has run a clean head-to-head test of AI against a qualified orthotist or prosthetist. Our evidence grade for Is it better than a person? is D, which is the grade we use when the comparison has not been measured, so we publish no parity number for this job at all.
What would move that grade is specific and testable. A trial comparing automated socket design against a certified practitioner’s design, judged on fit, skin integrity, and how often the device needs rework. A study of scan-to-fit pipelines that reports how many patients needed a clinician to re-cast by hand. Outcome data on 3D-printed orthoses fitted without a practitioner present. Until work like that exists, claims that software matches a trained fitter are assertions, not findings. We publish our grades and the reasoning behind them on the methodology page, and the underlying figures in the open dataset.
Good to know: 3D printing and digital scanning have changed how devices are made without removing the person who decides what to make.
When the picture could change
Most likely after 2042 (8 in 10 of our scenarios). What that window counts, and why we give a span instead of a date, is explained in how we estimate the replacement year.
Two things could pull it earlier. Cheap, accurate body scanning plus print-on-demand fabrication would cut the hands-on bench time per device. And software costs for the task slice AI can touch are very low next to the labor cost of the same hours, so clinics have a reason to automate the desk work fast.
Two things hold it back. The physical share of this job needs hardware at the dexterous humanoid tier, which is the hardest class of robot to build and the furthest from clinic-ready. And the work is licensed and clinically accountable: a fitting that causes skin breakdown is a patient-safety event, so a certified practitioner stays in the room even when machines do more of the making. For more on how hardware limits shape physical jobs, see our guide to humanoid robots and physical work.
How to stay needed in prosthetics and orthotics
Lean into the tasks the job keeps. Build a reputation for difficult fittings, the cases where standard sockets fail and someone has to solve it at the bench. Own the follow-up: repairs, modifications, and the review visits where a device that fit in March stops fitting in September. And get good at patient training, including the family members who do the daily care.
Two skills are worth time. Learn digital workflow properly, scan capture, CAD edits, and print tolerances, so you supervise the tools instead of waiting on a technician. Then learn to read and challenge outcome data, so you can say why one design choice beat another rather than guessing.
Close jobs are worth a look if you are weighing options. Opticians, dispensing and hearing aid specialists share the measure-fit-adjust pattern with a different device, and physical therapists work with many of the same patients. You can also browse the wider health technologists and technicians family or the healthcare sector.
To see where this role sits against others, put it beside another job on the compare tool, or scan the jobs that mostly need a person list.