Why the work stays beside the patient
This job is half pharmacy, half imaging. A technologist prepares and administers a radiopharmaceutical, then tracks where it travels in the body with a gamma camera or PET scanner. Software can count photons and clean up an image. It cannot draw up a dose, check a wristband, warm a cold arm or find a vein on a dehydrated patient.
The scan itself is a negotiation. Patients have to lie still for a long time, sometimes in pain, sometimes claustrophobic, sometimes very young. Positioning gets adjusted by hand and by eye. A technologist notices when breathing drifts, when a shoulder slips out of frame, or when someone is about to move, and fixes it before the data is wasted.
Then there is accountability for radioactive material. Doses are calculated, verified, logged and disposed of under state and federal rules, with a named person responsible. That responsibility is why much of the shift still belongs to a licensed human, even where the computer is doing the arithmetic alongside them. The job is also small and well paid: about 17,080 people in the US, with median pay of $101,370 and projected employment growth of 4.4% from 2025 to 2035 (BLS, 2025).
What AI does, what it helps with, what it leaves to people
The share of task time AI can handle on its own comes to 7% in our split. That part is the screen-shaped work: reconstructing and processing acquired images, and keeping the records and scan parameters that follow each study. These steps have clean inputs and a checkable output, which is where current models are strongest. Our coverage score for this occupation, the measure of task time AI can handle today, sits at 16 out of 100.
A further 20% of task time is assisted rather than taken over. Dose calculation and quality-control checks on the camera are good examples: the software proposes, the technologist confirms, and the signature stays human. Scheduling and protocol selection fall here too, because the plan still has to survive contact with a real patient’s condition.
That leaves 73% of the work with people. Administering radiopharmaceuticals, positioning the patient and monitoring them through the study are the clearest cases. So is explaining the procedure to someone who is frightened of the word nuclear. The physical side of the role lands in the mobile robot tier on the robotics row above, which is the expensive, slow-moving end of automation rather than a software update.
What the evidence actually shows
Our evidence grade for quality parity is D. In plain terms: there is no published head-to-head test of AI against nuclear medicine technologists on their own tasks, so we give no parity number at all. Most imaging AI research measures reading and reconstruction, which is the physician’s side of the department, not the technologist’s hands.
What would settle it is specific. A timed, audited trial in which an automated system prepares and administers a dose, sets up a real patient and completes an acquisition, scored against technologists for safety, repeat rates and image quality. Until something like that is published, the honest answer is that the hands-on work has not been measured. Our scoring method explains why an ungraded claim never becomes a number here.
Good to know: in diagnostic imaging, better software has mostly raised scan volumes, which keeps the people who run the scanners busy.
When the picture could change
Most likely after 2042 (8 in 10 of our scenarios). The replacement-year method sets out how that window is built and what it does and does not claim.
Two things could pull it earlier. Hybrid PET/CT suites that automate patient setup and table positioning would take a real physical task off the technologist. So would unit-dose radiopharmaceuticals delivered ready to inject, which shrinks the preparation step. Two things push it later. Radiation safety rules require an accountable licensed person for handling and administration, and that is a legal blocker, not a technical one. Capital cost is the other: a hospital replaces a camera over a decade or more, and the cost rows above show how far a year of staffing sits from a year of software in this department.
How to stay needed in nuclear medicine imaging
Lean into the parts of the shift that nobody is automating soon. Dose administration and venous access, patient positioning and in-study monitoring, and the safety and contamination response that follows a spill or a difficult injection. Those are the tasks the list above keeps with people.
Two skills compound. First, cross-modality competence, so you can cover CT on a hybrid scanner rather than only the camera. Second, the ability to audit a tool’s output: knowing when a reconstruction has smoothed away something real, and documenting it, is the skill that keeps you in the room when the software is doing more of the processing.
If you are weighing other imaging paths, it helps to read neighbors side by side. Compare radiologic technologists, magnetic resonance imaging technologists and radiation therapists, or put any two of them together on our compare tool.
What to do: look at the health technologists and technicians family and the wider healthcare sector page to see where this role sits, then check the list of jobs that mostly need a person for the pattern these roles share.