Why this work stays in the treatment room
Chiropractic care is a contact job. The center of the day is the adjustment itself: feeling how a joint moves under your hands, positioning the patient, then changing force and angle mid-motion based on what you feel and what the person says. Nothing about that step is text or data. It is touch, timing and a judgment call made in a second.
The hands-on exam sits in the same place. Palpating the spine, checking range of motion and screening for the signs that mean you should not adjust at all are decisions made with a body in front of you. So is the consent conversation that comes before treatment, where a patient asks what will happen and how much it will hurt.
Software is real in these practices, though. It is in the notes, the codes, the reminders and the intake forms. That is why the honest story here is task erosion rather than a job disappearing: the paperwork around care is shrinking faster than the care. Our coverage figure, which tracks the share of task time AI can handle today, reads 26 on the scale explained in how coverage is scored.
What software does, what it assists, and what people keep
AI handles a slice of the week on its own: 0% of task time. That is the clerical layer. Drafting visit notes from dictation, suggesting billing and coding entries, answering routine scheduling questions and sending recall messages are all jobs that assistants now do with light review.
A second group is assisted work, at 51% of task time. Here the tool prepares and the clinician decides. Imaging and intake summaries can be flagged for a second look. Home exercise plans and progress tracking can be generated, then edited to match what the patient can actually tolerate.
The rest, 49% of task time, stays with the person in the room. Manual manipulation and the hands-on physical exam are the two clearest examples. Both need pressure, feedback and a licensed professional who carries responsibility for the outcome.
How strong is the evidence?
Thin, and the page says so plainly. The evidence grade for this job is D, which means no study has tested an AI system against licensed chiropractors on this job’s real tasks. So we publish no parity number for it. A grade is a statement about measurement, not a prediction.
Two kinds of work would move it. First, a benchmark where AI assessments of musculoskeletal cases are scored blind against chiropractors’ assessments, including the decision not to treat. Second, a clinical trial comparing instrument-assisted or device-delivered manipulation with manual treatment on pain, function and adverse events. Until something like that exists, claims in either direction are opinion. You can read how we grade and date everything on the methodology page.
The labor market numbers are firmer. BLS counts about 39,630 chiropractors employed in the United States, with median pay near $79,200 and projected employment growth of 8.7% between 2025 and 2035 (BLS, 2025). That is a small, mostly self-employed occupation growing at a steady pace, which matters more for this job’s near-term outlook than any demo video.
When the picture could shift
Most likely after 2041 (8 in 10 of our scenarios). That window is a modeled median with an 80% spread, and the replacement-year method explains what it does and does not measure.
Two things could pull it earlier. Documentation and front-desk tools keep improving, which cuts the non-clinical hours a practice needs and reduces entry-level support roles before it touches clinicians. And instrument-assisted adjusting plus remote exercise programs could move some visits toward protocols that need less hands-on time.
Two things hold it back. The physical portion of this job falls in the dexterous humanoid tier on our robotics scale, and hardware that can safely palpate and manipulate a human spine is not a commercial product. Licensure, consent and liability sit on top of that: an adjustment is a regulated clinical act performed by a named licensee, and insurers and state boards move slowly. For a wider view of that hardware gap, see our guide to humanoid robots and physical jobs.
How to stay needed
Lean into the three tasks that the task list above keeps with people. Manual treatment is the obvious one, and the more varied and difficult the cases you can handle, the less a protocol substitutes for you. The hands-on exam and safety screen is the second: knowing when to refer out is a judgment patients and physicians both pay for. Third is the patient conversation, from consent through adherence coaching, which is where most of the repeat business actually comes from.
Two skills are worth adding. One is editing machine-drafted records well, so notes and codes are accurate and defensible rather than fluent and wrong. The other is behavior-change coaching, because home exercise compliance drives outcomes and an app alone rarely delivers it.
What to do: spend a month reviewing every AI-drafted note before it is filed, and keep a list of the errors you catch.
If you want to see how neighboring hands-on roles compare, look at physical therapists, physical medicine and rehabilitation physicians and acupuncturists. Wider context sits on the diagnosing and treating practitioners family page and the healthcare sector page. From there you can put two of these jobs side by side on compare any two jobs, or scan the full list in the job rankings.