Physical therapist

Why it resists AI

Physical therapy is about as automation-proof as a knowledge-adjacent career gets, and it earns the highest score in this batch. The work is hands-on by definition: manual therapy, joint mobilization, guiding and physically supporting patients through exercises, and reading a body's response in real time. The BLS lists dexterity ("use their hands to provide manual therapy … comfortable massaging and otherwise physically assisting patients") and physical stamina among required qualities. This is non-routine physical work performed on unpredictable, living subjects, the exact category Frey & Osborne's framework rates least automatable.

The cognitive part of the job is also resistant. A PT reads a medical history, watches how a specific person stands and walks, diagnoses a functional movement problem, and builds a plan that they then adjust week to week based on hands-on feedback. The "GPTs are GPTs" exposure work places this kind of in-person, individualized clinical judgment far from the high-exposure desk occupations. AI can suggest exercise protocols or analyze gait video, but it can't perform the manual treatment, motivate a patient through pain, or carry the legal responsibility for the plan.

The licensing moat is firm. All states require PTs to be licensed, which means earning a Doctor of Physical Therapy degree and passing the national licensure exam, with continuing education to maintain it. Add the highest projected growth in this group (11%), and every signal in the methodology points the same direction.

What the work is actually like

PTs are on their feet all day working with patients, and because they often lift and reposition people, the BLS flags back injury as a real occupational risk that good body mechanics only partly mitigate. It's physical, but it's also relational. You see the same patients repeatedly over weeks, build trust, and coach them through slow, sometimes frustrating recovery. "Compassion" and "communication skills" are listed as core required qualities.

Settings vary. About 34% of PTs work in dedicated therapy offices and clinics, 28% in hospitals, 11% in home health (median pay is actually highest here), plus nursing facilities and self-employment. Most work full time during normal business hours, though some evenings and weekends exist, and part-time work is common. That's a real contrast to the round-the-clock shift demands of hospital nursing.

Pay and earning trajectory

The median annual wage was $101,020 in May 2024, the highest of the four healthcare roles here. The lowest 10% earned under $74,420, a notably high floor, and the top 10% over $132,500. By setting, home healthcare led at $108,110, followed by nursing and residential care ($105,330) and hospitals ($105,140).

Earnings rise with experience and specialization. Once they have experience, PTs can pursue board certification in areas like orthopedics, sports, or geriatrics (via exam plus clinical work), which can support higher pay and referral demand. Some open or buy into private practices. The ceiling is solid. The up-front educational cost is the real catch (below).

How to get there from tech

This is the longest road of the four. Becoming a PT requires a Doctor of Physical Therapy (DPT), a doctoral degree that itself takes about 3 years, and DPT programs require a bachelor's plus prerequisite coursework (anatomy, chemistry, physics) for admission. If your tech bachelor's didn't include those sciences, plan on a year or more of prerequisites first, which makes the realistic total roughly 5 to 7 years end to end.

Cost is the honest sticking point. DPT tuition and resulting debt vary enormously by school. Public-program tuition can land near $42k/year, while reported average graduate debt ranges from roughly $70k to over $150k depending on the program. Then you sit for the national licensure exam and meet state requirements (some add a law exam and background check).

What transfers from tech: the heavy science prerequisites and biomechanics and neuroscience coursework reward analytical, methodical learners, and documentation and EHR fluency carries over. But the core of the job, manual treatment and in-person coaching, is built almost entirely in supervised clinical rotations.

Honest tradeoffs vs. a tech job

Outlook & demand

BLS projects 11% growth from 2024 to 2034 ("much faster than average"), with about 13,200 openings per year. The demand drivers are durable: an aging population with more strokes, mobility injuries, and chronic musculoskeletal conditions, plus a healthcare-wide shift toward non-opioid pain management that leans directly on physical therapy. This is structural demand, backed by demographics, that automation is not positioned to erode.