■ HIGH RISK ■ Healthcare
The job survives, but it's being deskilled around the edges — AI now assists with positioning, exposure, and first-pass reads. Patient handling keeps techs employed; the technical judgment that made the role skilled is what's eroding.
“Automated positioning and AI reads. You press the button and say 'hold still.'”
Our AI replacement risk score — how we score jobs
An X-ray tech's shift is a conveyor of humans who don't want to be there: verify the order, check pregnancy status, position the patient against the bucky, set technique factors, shoot, review the image for rotation or clipped anatomy, repeat or release. Between patients there's QA on the equipment, radiation-safety protocol, and a steady stream of portable exams on the floors — shooting chest films around IV lines on patients who can't sit up, which is where the actual skill lives.
Modern rooms are automating the technical decisions one by one. Automatic exposure control has handled dose for years; now camera-guided systems suggest positioning, auto-collimate, and flag a clipped costophrenic angle before the radiologist ever sees it. AI triage tools pre-read for pneumothorax and fractures, and repeat-rate analytics grade the tech's work continuously. None of this removes the human from the room — someone still has to move a 90-year-old hip fracture patient onto the table without screaming — but it compresses the expertise gap between a ten-year tech and a ten-month one. That compression is what a 65 risk score looks like in practice: same headcount pressure, lower skill premium, and imaging volumes that keep techs busy without making them irreplaceable.
What holds firm is everything that involves touching people. Trauma patients, pediatric patients, combative patients, portables in the ICU — no positioning algorithm handles a human who is frightened, in pain, or attached to six lines. Techs who cross-train into CT, MRI, or interventional radiology climb away from the most automatable work. The ones doing routine chest and extremity films in an outpatient clinic are standing closest to the machinery.
Automatability: our editorial assessment of current and near-term AI capability
Pressure builds steadily through 2030. AI-assisted positioning and auto-QA are shipping in new rooms now, and imaging volumes are growing slower than automation is improving. Expect flat-to-declining demand for techs doing routine outpatient radiography by decade's end, while CT, MRI, and interventional roles stay tight. The job won't vanish — patients still need handling — but the skill premium for plain-film expertise erodes visibly within five years.
It's a decent entry point but a risky place to stop. Routine radiography is the most automatable corner of medical imaging, and AI-assisted rooms are narrowing the skill gap. The techs who thrive treat plain film as a launchpad into CT, MRI, or interventional work, where equipment complexity and patient acuity keep humans essential far longer.
AI already pre-reads for specific findings — pneumothorax, fractures, tuberculosis screening — and radiologists increasingly work with an algorithmic first pass. But that's the radiologist's problem more than the tech's. For techs, the nearer threat is automated positioning and quality control, which shrinks the expertise that separated great techs from adequate ones.
Interventional radiology and cath lab work top the list — real-time, hands-on, high-stakes. MRI is strong because of protocol complexity and safety screening. CT sits in the middle. Routine outpatient plain film is the most exposed. Mammography retains regulatory human requirements. Generally: the sicker the patient and the more complex the machine, the safer the job.
Stack credentials — a second or third modality registry is worth more than years of seniority in plain film. Volunteer for the hardest assignments, because portable and trauma skills are automation-proof. And get hands-on with the AI tools your department buys; being the tech who validates the algorithm's output is a promotion, not a demotion.