Why this work stays in human hands
Massage therapy is paid touch. The core of the job is kneading muscle and soft tissue, reading how a body responds under your hands, and changing pressure in the same second you feel resistance. That loop of feeling and adjusting is the service, not an add-on to it. So the honest answer to whether AI will replace massage therapists is that the hands-on hours stay with people, while the paperwork around them thins out.
Two tasks show why. Assessing soft tissue condition, joint function, muscle strength and range of motion is done by palpation, by small questions, and by watching a client wince or relax. Conferring with clients about their medical history and current problem shapes everything that follows, including the decision not to work an area at all. A machine can follow a mapped route across a back. It does not own the judgment call to stop.
There is also the part clients rarely name out loud. People book massage for relief and for being cared for by someone who is present. Robotic massage tables now sell on their consistency and their price, and spas and gyms are testing them. Consistency is useful for a tight hamstring after a run. It is a different product from an hour with a licensed therapist who knows your shoulder history.
Our data puts the human share of task time at 77% for this occupation, and the overall coverage score, which answers whether AI can do the work today, at 10 out of 100. You can read how that figure is built on the coverage method page.
What software handles, what it assists, and what it leaves alone
Software already carries the desk side of a practice. Maintaining client treatment records and intake forms is routine documentation, and scheduling, rebooking and reminders are the same. Across the share of task time AI could touch, the part it could handle on its own and the part where it only assists come out close to even: 0% versus 23%.
The assist column is where most therapists will meet AI first. Drafting a proposed treatment plan from intake notes, or turning a session into plain stretching, strengthening and relaxation guidance a client can actually follow at home, are both faster with a tool than with a blank page. A therapist still signs off, because the plan has to match what the tissue did on the table.
What is left alone is the hour itself. Massaging and kneading muscle, applying finger and hand pressure to specific points, and judging when to refer a client on to another health professional all sit in the needs-a-human group. The robotics side explains part of that: this is dexterous whole-body work with a real person underneath, which is the hardest tier of physical automation, not the easiest.
What the evidence actually shows
Our quality-parity grade for massage therapists is D. That grade means there is no published head-to-head test of AI or a massage robot against a licensed therapist on this job’s real tasks, so we publish no parity number for it. Marketing claims from device makers are not tests, and neither is a positive review.
What would settle it is specific and not hard to imagine: a trial comparing robotic and therapist-delivered sessions on pain and range-of-motion outcomes, with the same clients, over weeks rather than one visit, plus safety data on contraindications a machine has to catch. Until something like that is published and repeatable, the parity question stays open rather than answered. How grades are assigned is set out on the quality-parity method page, and the wider approach is on the methodology page.
The labor market data is firmer. BLS counts about 98,790 massage therapists employed in the United States, with median pay of $58,450 and projected employment growth of 15.3% from 2025 to 2035 (BLS, 2025). That is a growing occupation, not a shrinking one, though much of it is self-employed or contract work where your own client book matters more than any hiring trend.
When the picture could change
Our replacement-year estimate for this occupation reads as follows. Most likely after 2045 (8 in 10 of our scenarios). What that window measures, and how the spread is produced, is explained on the replacement-year method page.
Two things could pull it earlier. First, cheaper hardware: per-session machine costs are far below the cost of therapist hours, so high-volume venues like gyms, airports and hotels have a clear reason to install tables. Second, real progress in dexterous manipulation and force sensing, the same engineering that would let a robot handle soft, variable, living material safely.
Two things hold it back. Licensing and liability sit on a person, and no state licenses a machine to decide that a client’s symptom needs a doctor instead of deep tissue work. And demand is partly for the relationship: regulars rebook with a named therapist, and that preference is hard to automate away.
Good to know: robotic tables are most likely to take the quick, generic sessions first, which is where unlicensed chair and franchise work already sits.
How to stay needed
Lean into the tasks that stay human. Get better at assessment, so your hands and questions find the cause rather than the sore spot. Build clinical treatment planning you can defend to a physical therapist or physician. And treat referral judgment as a skill, not a formality, because knowing when not to treat is what a machine cannot own.
Two skills are worth time: a clinical specialty with medical referral paths, such as injury rehab, oncology or prenatal work, and simple business and marketing habits so your book fills from repeat clients rather than discount platforms. Using AI tools for notes and client follow-ups is fine, and it buys you hands-on hours.
If you want to see how close occupations compare, three are worth a look: physical therapist aides, physical therapist assistants and occupational therapy assistants. You can also read the rest of the other healthcare support occupations, the wider healthcare sector, or the list of jobs that mostly need a person. To set this job beside another one, use compare any two jobs, or check the headline figure of 83 out of 100 (higher is safer) against the full job rankings.