Pivotum/All careers/Allied Health/Fall 2026

Is Physical Therapy Safe From AI?

Free sampler. Re-scored every six months.

The findingExposureProtectionMethod
2.5AI exposurewhere 10 is most at risk
The short answer

Hands-on physical therapy scores 2.5 out of 10 for AI exposure, where 10 is most at risk — the lowest score of any career in this index apart from surgery. Occupational therapy scores 2.6, speech-language pathology 3.1. The screen-based corner of allied health scores considerably higher. This is what maximum protection looks like in our framework: bodies, trust, licensure and unpredictability all at once.

The short answer for parents: yes, and it is barely a close call. The risk question here is effectively settled. The real questions are the traditional ones — the cost of the doctorate against the salary, and appetite for physical work over three or four decades.


Allied health AI risk score by role

Every career in this index is scored 1–10, where 10 is most exposed to AI. Same six factors, same weights, applied identically to a therapist and a paralegal.

Allied Health track20232025Now3-yr moveBand
Physical therapy (hands-on)2.02.22.5+0.5Very low
Occupational therapy2.12.32.6+0.5Very low
Speech-language pathology2.62.83.1+0.5Low
Respiratory therapy2.62.83.1+0.5Low
Diagnostic imaging technologist3.54.04.4+0.9Low–Mod
Records / screen-based allied roles5.15.66.1+1.0Mod–High

2023 and 2025 figures are reconstructed using current methodology, not archived from past editions.

For scalethe median career in this edition scores around 5.5. Bedside nursing scores 2.8. Surgery scores 2.1. Entry-level software development scores 8.1.

The +0.5 on hands-on therapy is all administrative. Documentation and exercise-plan drafting. The clinical work has not moved at all.


Will AI replace physical therapists?

Almost nothing about the core of this job is automatable with anything on the visible horizon.

AI is takingIt can't touch
Documentation and progress notesHands on a body
Exercise plan draftingWatching a gait in person
Scheduling and billingFeeling resistance, guarding, compensation
Outcome measure scoringMotivating someone through months of painful work
Insurance authorization paperworkAdjusting in real time to what the body does
Home exercise remindersKnowing when to push and when to stop

Demand is also moving independently of technology: physical therapists are projected to grow 11% and occupational therapists 14% through 2034, both far faster than average, driven by an aging population and rising survival rates after serious illness.


Why does it score so low? The six factors

How hands-on physical therapy rates against each. Ratings are 0–10 on each factor's own terms.

How much of this job can AI already do?4.9
How hard will it be to land that first job?2.0
Does it have to be done in person, with your hands?9.5
Does someone need a human they can trust and hold responsible?9.0
Does the law require a licensed human?9.0
How often does the job hit genuinely new, high-stakes situations?8.5

One factor, worked through in full

So you can see what the analysis actually looks like.

Does it have to be done in person, with your hands? — rated 9.5

This is the highest physical rating we assign to any career, and it is doing most of the work in a score of 2.5. Therapy is manual by definition: hands on a body, resistance felt rather than measured, adjustment in real time to how a person actually moves.

Robotics lags software badly, and it lags furthest in exactly this kind of setting — an unpredictable human body, in a room, over months.

But of our four protection factors, physical is the one we rank as least durable over a long horizon, and this profile is where that caveat matters most. A seventeen-year-old entering a doctoral program will practice for forty years. Licensure and human trust move at the speed of law and society and are unlikely to shift much. Physical protection moves at the speed of robotics, which is the one thing actively targeting it.

There is already early evidence worth taking seriously. Studies of robotic-assisted rehabilitation report meaningfully better motor outcomes in some domains than conventional therapy — not merely comparable, better. If robotic delivery is superior for certain protocols, then "hands-on work is safe" starts looking less like a wall and more like a line.

Two things keep our score where it is. Better outcomes with robotic assistance is not the same as therapy delivered without a therapist — in every deployment we found, a licensed clinician prescribes, supervises and adapts. And the peer-reviewed literature is explicit that translation into routine practice remains limited.

The thing to watch is a specific transition, and it is the same pattern that splits the skilled trades: whether robotic systems move from assisting a therapist to substituting for therapist hours in structured, protocol-driven rehabilitation. Structured settings are where automation always arrives first. Unstructured ones — a stroke patient in their own home, an athlete with an atypical injury — are last.


Which allied health roles are safest from AI?

Ranked by exposure, safest first:

  1. Physical therapy — 2.5. Maximum physical involvement, licensure, trust and individual variation.
  2. Occupational therapy — 2.6. The same protections applied to daily function.
  3. Speech-language pathology — 3.1. Deeply relational, though less physically manual.
  4. Respiratory therapy — 3.1. Physical and acute, with more protocol-driven content.
  5. Diagnostic imaging technologist — 4.4. Patient positioning is physical; image analysis is not.
  6. Records / screen-based allied roles — 6.1. Same field, no protection.

The pattern holds across every profession we score: move the same work behind a screen and exposure roughly doubles.


Is physical therapy a good career in 2026?

The AI question is about as settled as this index gets. The real considerations are unchanged.

The doctorate is expensive and the salary, while solid, has not kept pace with the cost of the training in many markets — that is the most common complaint from people in the field, and it has nothing to do with technology.

The work is physically demanding over decades. Some patients do not improve, and sitting with that is harder than students expect.

Against that: demographics are a genuine and durable tailwind, the work produces visible progress in a way few jobs do, and the relationships run over weeks and months rather than minutes.

Within allied health more broadly, the same rule applies everywhere: protection lives in the hands-on roles and thins in the screen-based ones.


Common questions

Will AI replace physical therapists?
Not on any visible horizon. The core work is manual, individual and adaptive — the hardest combination for automation. Documentation is automating, which most therapists welcome.
Will robots replace physiotherapists?
This is the more serious version of the question. Robotic-assisted rehabilitation is real and in some protocols produces better outcomes. So far it assists rather than substitutes, and a licensed clinician still prescribes and supervises. It is the thing we watch most closely in this profile.
Is physical therapy a good career?
On AI exposure, one of the best available. The genuine trade-offs are the cost of the doctorate and physical demands over a long career.
Is PT safer than nursing?
Marginally — 2.5 against 2.8. Both are extremely well protected, and both have legally mandated clinical training that protects the entry route.
What about occupational therapy or speech therapy?
Both score very well — 2.6 and 3.1. Speech-language pathology is consistently under-recruited, and OT assistant roles are projected to grow strongly with a shorter training route.

Related profiles


What's in the full allied health profile

This sampler tells you where allied health stands. The full profile tells you what to do about it.

FreeFull
Verdict, all sub-track scores, 3-year trend
Six-factor ratings
Reasoning behind every factor ratingone example
How durable each protection is — where AI is already pressing
The honest downsides
What's genuinely good about it — satisfaction data
Who this work suits, and who it doesn't
The AI-native advantage — how to prepare
Routes in
Where the degree leads later — and which exits raise exposure
Program evaluation checklist
Questions to ask an admissions office — twelve, plus red flags
Sourced further reading, including the strongest case against our score
Discussion questions for parent and student
A short version written directly to the student
Technical scoring appendix

Get the full profiles

Most families are weighing two or three careers seriously, and a few more they haven’t ruled out. Pick the ones you need.

Each includes a short version written directly to the student and the technical scoring appendix. Spring 2027 updates of whatever you buy are included.

Read one complete profile free → We publish computer science in full so you can judge the depth before buying anything.

28 careers, scored the same way. Scores measure exposure to what AI can already do — not how much any particular employer has deployed.
2023 and 2025 figures are reconstructed using current methodology, not archived from past editions.
Re-scored every six months. We publish where we might be wrong.
Analysis and scoring judgments are ours. Drafting is AI-assisted — how this is written.
Terms · Privacy · Refunds