SAFE
The core of this job is physical: positioning an anxious, immobile, or injured patient on a table, shielding them, adjusting tube angle and technique factors, and getting a diagnostically usable image on the first try — none of which language models touch and none of which current robotics can do in a busy ED. AI is genuinely eating the adjacent layers: protocol selection, dose optimization suggestions, image quality flagging, and the radiologist's read are all being automated, which compresses the tech's decision space rather than the tech's body. Certification (ARRT) and state licensure are required in most states, and radiation-safety accountability sits with a named human.
Headcount grew steadily across the period.
Median pay $60,510 → $80,110 +5.9% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+4.3% 228,000 → 237,800 on the projections basis
Hard to automate, and growing
The work resists current AI and the BLS projects +4.3% more of these jobs by 2034. Note that safe does not mean well paid — several of the fastest-growing resistant occupations are among the lowest paid on the register.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~12,900 openings a year on average, including replacing people who leave.
SkiagrapherMammographerRadiographerRadiology AideX Ray Operator3D TechnologistImaging SpecialistImaging TechnicianRadiology AssistantRadiology TherapistRadiation TechnicianRadiology SpecialistMammography TechnicianRadiography TechnicianRadiological TechnicianRegistered RadiographerInterventional TechnicianRadiographer TechnologistRadiological TechnologistMedical Imaging TechnicianDocument Imaging TechnicianMulti-Modality TechnologistRadiology Technologist (RT)Limited Radiology Technician
Holding it up: embodiment . Weakest point: judgment & accountability .
Tasks largely resist digitisation Positioning a hip-fracture patient for a cross-table lateral, or coaxing a trauma patient into a swimmer's view when they cannot lift an arm, is improvised manual work per body habitus and injury — the digitisable parts (AEC, DR post-processing, protocol presets, dose tracking) have already been automated, which is why this is 15 and not 18.
Hands-on in uncontrolled environments You are in the room, moving portable units through ED bays and ICUs, transferring patients, palpating landmarks, and wearing lead while doing it — an uncontrolled environment with an immobile, bleeding, or combative subject, which is about as embodied as diagnostic imaging gets short of surgery.
Licensed human required and personally liable ARRT certification plus state licensure in most states means your credential number attaches to the exam and to the dose delivered, and a state radiation-control board can sanction you personally for a repeat-heavy or unshielded exam — but the diagnostic call and the order sit with a physician, so you carry procedural accountability rather than the interpretive liability that would push this past 15.
Meaningful discretion You decide when to repeat, when a patient cannot hold still enough to justify the dose, and when the ordered view will not answer the clinical question and radiology needs a call — genuine discretion, but bounded by written protocols, the radiologist's order, and ACR technique standards, which is why this sits at 10 and not in the teens.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (15/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (13/20) is whether the law requires a licensed human to sign. Trust premium (11/20) is whether buyers specifically pay for a person. Judgment and accountability (10/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 34 of this occupation's 67 points (51%).
Embodiment (18/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 80/100, still SAFE.
Task-mix shift is real here and has two tiers: if AI takes protocol selection and image-quality flagging, what remains is the hard tier — trauma/pediatric/bariatric positioning, non-cooperative patients, portable exams in the ICU, contrast reaction recognition and first response. Formal recognition of this via advanced-practice credentials (ARRT Radiologist Assistant / RPA scope expansion, already licensed in a handful of states) would put consequential calls under ambiguity formally on the tech.
Closing the state licensure gap: as of now roughly a dozen states (e.g., FL/TX have licensure, but states like Idaho, Missouri, North Carolina lack general radiographer licensure). Passage of the federal CARE Act (Consistency, Accuracy, Responsibility and Excellence in Medical Imaging and Radiation Therapy Act, reintroduced in multiple Congresses) or state-by-state licensure bills would make a named licensed operator legally mandatory nationwide, plus NRC/Agreement-State rules naming the technologist as the individual accountable for dose delivered under an AI-generated protocol.
A Joint Commission or ACR accreditation requirement that an AI-suggested dose/protocol modification be accepted and initialed by the credentialed technologist before exposure, rather than auto-applied by the scanner — the imaging analogue of pharmacist countersignature. ACR AI-QA registries and the FDA's push on 'human in the loop' for autonomous imaging devices are the visible precursors.
Same two-tier mechanism: automation of throughput and QC work leaves the residual day weighted toward physically unrepeatable exams and patient management. Watch for hospital staffing models that pool 'routine outpatient' scans under autonomous positioning aids while keeping ED/OR/ICU techs — the remaining role is more resistant than the average one today.
The limit. trust_premium has no plausible route: patients do not choose or pay for a specific technologist, the exam is ordered by a physician and billed as a facility service, and no buyer-side preference mechanism exists. Do not expect movement there.
| New York-Newark-Jersey City, NY-NJ | 12,700 | $102,730 +28% |
| Los Angeles-Long Beach-Anaheim, CA | 6,450 | $108,030 +35% |
| Chicago-Naperville-Elgin, IL-IN | 5,650 | $82,140 +3% |
| Dallas-Fort Worth-Arlington, TX | 5,470 | $83,140 +4% |
| Houston-Pasadena-The Woodlands, TX | 5,000 | $80,780 +1% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 4,630 | $78,890 -2% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 4,130 | $80,190 +0% |
| Boston-Cambridge-Newton, MA-NH | 4,100 | $104,590 +31% |
| Vallejo, CA | 270 | $160,380 +100% |
| San Jose-Sunnyvale-Santa Clara, CA | 1,000 | $154,850 +93% |
| San Francisco-Oakland-Fremont, CA | 2,850 | $141,780 +77% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 67. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.