Radiation Therapists

29-1124.00
Median wage $105,310/yr17,070 employed (US)Rank #813 of 923 scored · top 88% by substitution

Provide radiation therapy to patients as prescribed by a radiation oncologist according to established practices and standards. Duties may include reviewing prescription and diagnosis; acting as liaison with physician and supportive care personnel; preparing equipment, such as immobilization, treatment, and protection devices; and maintaining records, reports, and files. May assist in dosimetry procedures and tumor localization.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution14
Exposure14
Augmentation46

Substitution — the headline: capability discounted by cost, barriers and adoption.

Exposure — technical capability alone, regardless of whether anyone deploys it.

Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.

Tasks on the substitution scale

22 rated tasks, binned by substitution score.

Position among all scored occupations

Distribution of 923 occupation scores; the marker is this occupation.

Tasks with substitution ≥ 70

5%

Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.

Why this score

The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.

Task automatabilityw 35%14

panel mean rating 1.6/5 → substitution pressure 14/100

Technical feasibility todayw 20%14

panel mean rating 1.5/5 → substitution pressure 14/100

Cost vs. human wagew 15%15

panel mean rating 1.6/5 → substitution pressure 15/100

Adoption barriersw 20%inverted — strong barriers lower the score11

panel mean rating 4.6/5 (barrier strength) → substitution pressure 11/100

Sector adoption velocityw 10%20

panel mean rating 1.8/5 → substitution pressure 20/100

Task breakdown (22 tasks)

Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.

Schedule patients for treatment times.

76

CI 72–79 · exposure 75 · augmentation 75 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Healthcare organizations have widely adopted electronic scheduling systems and some AI-driven optimization for patient flow; the healthcare information technology sector shows strong and ongoing adoption of automation in administrative tasks like scheduling.
Sector adoption velocityclaude-sonnet-53/5Healthcare administrative functions are adopting scheduling automation at a moderate pace, slower than pure information sectors due to EHR integration complexity and clinical workflow dependencies.
Augmentation potentialclaude-haiku-4-5-202510014/5AI scheduling systems augment human schedulers by instantly showing optimal slots, flagging conflicts, suggesting batch times, and automating routine bookings, allowing schedulers to focus on complex cases and patient communication rather than manual calendar entry.
Augmentation potentialclaude-sonnet-54/5AI-assisted scheduling tools significantly reduce administrative burden on radiation therapists and staff by optimizing slots and reducing conflicts, while humans retain oversight of clinical prioritization.
Task automatabilityclaude-haiku-4-5-202510014/5Scheduling patients for radiation therapy treatment times is largely routine and rule-based, involving calendar management, availability matching, and constraint satisfaction. Current AI can handle this end-to-end through calendar APIs, patient databases, and scheduling algorithms, achieving >50% time savings with proper integration, though some edge cases (complex medical conflicts, priority overrides) may still require human review.
Task automatabilityclaude-sonnet-54/5Scheduling patients for treatment slots is a well-structured, rules-based task involving calendar management, resource availability, and constraints that off-the-shelf scheduling software and AI-assisted systems handle well.
Adoption barriersclaude-haiku-4-5-202510012/5Few hard legal barriers exist to automating scheduling itself; however, patient communication, treatment-plan confirmation, and coordination with clinicians still typically require human sign-off, and some organizations prefer human-facing scheduling for patient experience and liability reasons.
Adoption barriersclaude-sonnet-52/5No licensure requirement specifically covers scheduling itself, though it is embedded within clinical workflows requiring coordination with treatment planning, creating some organizational friction.
Cost vs. human wageclaude-haiku-4-5-202510015/5Automated scheduling systems are inexpensive to operate per task once deployed—inference and database queries cost pennies compared to the loaded cost of a scheduler's hourly wage for the same function.
Cost vs. human wageclaude-sonnet-54/5Automated scheduling systems are far cheaper per transaction than having a clinical therapist or scheduler manually manage calendars, though integration and oversight costs exist.
Technical feasibility todayclaude-haiku-4-5-202510014/5Deployed scheduling software and healthcare management systems already automate much of patient scheduling in practice, including radiation oncology departments. Mature products integrate with existing hospital systems and demonstrate reliable performance, though integration complexity and medical-specific rules sometimes require manual oversight or adjustment.
Technical feasibility todayclaude-sonnet-54/5Healthcare scheduling software with automated/AI-assisted appointment optimization is widely deployed in radiation oncology departments today, though some manual coordination for complex treatment courses remains.

Maintain records, reports, or files as required, including such information as radiation dosages, equipment settings, or patients' reactions.

44

CI 36–51 · exposure 42 · augmentation 75 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Radiation oncology is a highly digitized, regulated specialty with mature EHR and RTPS adoption across major hospital systems and cancer centers. Automated record capture and dose tracking are standard practice in production environments, with rapid ongoing integration of AI-assisted documentation.
Sector adoption velocityclaude-sonnet-53/5Healthcare IT adoption is steady but slower than pure information-sector adoption due to regulatory, interoperability, and safety-critical constraints.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can assist radiation therapists by auto-populating dosage fields, flagging anomalies in treatment parameters, and summarizing patient reactions for faster charting. These augmentations reduce documentation burden while preserving human verification and clinical judgment.
Augmentation potentialclaude-sonnet-54/5AI-assisted documentation, auto-population of equipment settings, and templated reporting meaningfully speed up record-keeping while the therapist remains responsible for accuracy.
Task automatabilityclaude-haiku-4-5-202510012/5Record-keeping of structured data (dosages, settings) is automatable, but radiation therapy records involve clinical judgment about patient reactions and contextual notes that require human review and interpretation. Current systems can extract and organize structured fields but cannot reliably capture nuanced clinical observations without human oversight.
Task automatabilityclaude-sonnet-53/5Structured data logging and report generation from equipment outputs and clinical notes can be substantially automated, but reconciling patient reactions and clinical observations still requires human judgment and entry.
Adoption barriersclaude-haiku-4-5-202510014/5Radiation therapy records are subject to regulatory requirements (FDA, state licensure, HIPAA, accreditation standards) and must be legally auditable and traceable to a licensed professional. Clinical validation and signature requirements impose material legal and accountability barriers to full automation.
Adoption barriersclaude-sonnet-54/5Medical record accuracy for radiation dosage carries significant liability and regulatory documentation requirements (e.g., safety and compliance rules), requiring human verification and accountability.
Cost vs. human wageclaude-haiku-4-5-202510014/5Automated record systems (integrated EHR/RTPS) are substantially cheaper than staffing a person for full-time documentation and clerical work. The per-record cost of system maintenance and oversight is orders of magnitude lower than loaded labor costs for manual transcription and filing.
Cost vs. human wageclaude-sonnet-53/5Automated data capture reduces clerical time modestly, but the therapist must still verify and document clinically, so overall cost savings are moderate rather than dramatic.
Technical feasibility todayclaude-haiku-4-5-202510013/5EHR and radiation oncology software platforms have modules for capturing dosages and equipment settings with significant automation, but they still require human data entry, validation, and clinical interpretation. Systems exist but depend on structured templates and human oversight to ensure accuracy and compliance.
Technical feasibility todayclaude-sonnet-53/5EHR/oncology information systems already auto-capture dosage and equipment settings, but integration is uneven across facilities and free-text reaction notes still require manual entry and review.

Review prescription, diagnosis, patient chart, and identification.

24

CI 20–29 · exposure 25 · augmentation 50 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI remains cautious and highly regulated. While some hospitals pilot AI documentation tools, verification and sign-off of radiation therapy prescriptions remain rare in production, reflecting both regulatory conservatism and malpractice risk sensitivity.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially radiation oncology, adopts AI cautiously due to safety-critical workflows and regulatory oversight, with pilots for decision support but slow deployment of autonomous verification.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by automatically flagging inconsistencies between prescription and chart, surfacing prior diagnoses, and organizing patient data for review, thereby speeding the therapist's verification process while maintaining human oversight and sign-off.
Augmentation potentialclaude-sonnet-53/5AI can help surface inconsistencies between chart, prescription, and identification data, speeding up the therapist's review, though the therapist must still perform and confirm the check.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can extract and summarize text from charts and prescriptions, verifying medical accuracy and catching critical discrepancies requires human judgment and accountability. Current systems lack the clinical reasoning and liability tolerance to autonomously sign off on patient identification and prescription matching.
Task automatabilityclaude-sonnet-52/5AI can extract and cross-check structured data from records, but final verification of prescription, diagnosis, and patient identity for radiation treatment requires clinical judgment and accountability that current systems cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Radiation therapy is heavily regulated; a licensed radiation therapist must legally verify and sign off on prescription accuracy and patient identification before treatment. Liability for incorrect verification rests with credentialed staff, creating a hard legal and professional barrier to full automation.
Adoption barriersclaude-sonnet-55/5This is a licensed clinical safety check in radiation oncology; regulations and institutional protocols mandate a qualified radiation therapist to verify patient identity and treatment prescription before treatment.
Cost vs. human wageclaude-haiku-4-5-202510013/5AI-assisted document processing has become relatively inexpensive, but the cost of integration, training on hospital systems, and maintaining oversight infrastructure roughly balances the wage cost of a radiation therapist performing this verification task.
Cost vs. human wageclaude-sonnet-52/5AI-assisted chart review tools add licensing and integration costs while still requiring a human therapist to perform the actual verification, so cost savings versus the human-only step are modest.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some optical character recognition and data extraction products exist, but no mature systems reliably and independently review medical prescriptions against patient identification and diagnosis charts in production clinical settings. Error rates remain too high for unsupervised deployment.
Technical feasibility todayclaude-sonnet-52/5Some EHR/clinical decision-support tools flag discrepancies or assist chart review, but no deployed product autonomously performs this verification step reliably in production without a therapist's confirmation.

Help physicians, radiation oncologists, or clinical physicists to prepare physical or technical aspects of radiation treatment plans, using information about patient condition and anatomy.

24

CI 20–28 · exposure 25 · augmentation 63 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Adoption of AI-assisted tools in radiation oncology is slow and cautious; institutions are primarily exploring pilot implementations for segmentation and planning support rather than deploying autonomous systems. The field's high safety standards and regulatory scrutiny limit rapid uptake.
Sector adoption velocityclaude-sonnet-53/5Radiation oncology departments have adopted AI-assisted contouring and planning tools at a moderate pace, with pilots and some production use in academic and larger centers, though not yet universal.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by automating organ segmentation from imaging, suggesting initial dose distributions, or flagging anatomical anomalies, thereby accelerating plan preparation. However, the human physicist remains essential for validation, optimization, and clinical responsibility.
Augmentation potentialclaude-sonnet-54/5AI tools for auto-segmentation, dose prediction, and plan optimization meaningfully speed up the technical planning process while therapists, physicists, and physicians retain final review and decision authority.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in analyzing patient anatomy from imaging and suggest dosimetry parameters, the task requires extensive clinical judgment, patient-specific customization, and integration with physician oversight. Current systems cannot reliably prepare treatment plans end-to-end with the precision and liability tolerance required in clinical oncology.
Task automatabilityclaude-sonnet-52/5AI can assist with contouring and dose optimization suggestions, but the collaborative, judgment-heavy planning process with physicians and physicists still requires substantial human integration and cannot be fully automated end-to-end today.
Adoption barriersclaude-haiku-4-5-202510014/5Radiation therapy is heavily regulated (FDA oversight, state licensure requirements for physicists), and there are strict liability constraints around incorrect dose calculations or anatomical misinterpretation. Clinical physicists and oncologists must legally review and sign off on treatment plans, creating strong barriers to full automation.
Adoption barriersclaude-sonnet-55/5Radiation treatment planning is tightly regulated, requires licensed physicians, radiation oncologists, and certified medical physicists to sign off, and errors carry severe patient safety and liability consequences.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI systems into radiation therapy workflows requires significant validation infrastructure, physician oversight, and liability management. The cost of ensuring safety and compliance remains comparable to or exceeds the cost of expert human planning.
Cost vs. human wageclaude-sonnet-52/5AI planning tools reduce some contouring/optimization time but still require expensive clinical oversight, hardware, and software licensing, so overall cost savings versus therapist/physicist time are moderate rather than order-of-magnitude.
Technical feasibility todayclaude-haiku-4-5-202510012/5AI systems exist for segmentation and dose prediction in research settings, but few if any are deployed in production radiation therapy workflows as autonomous plan preparators. Most clinical applications remain in pilot or advisory mode, with human physicists validating all outputs.
Technical feasibility todayclaude-sonnet-52/5Deployed treatment planning systems include AI-assisted auto-contouring and knowledge-based planning tools, but these are narrow components integrated into a human-led workflow rather than performing the full collaborative planning task reliably.

Calculate actual treatment dosages delivered during each session.

24

CI 20–28 · exposure 30 · augmentation 75 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Hospitals and cancer centers adopt treatment planning software incrementally, but automation of dosage sign-off remains limited due to regulatory requirements and malpractice risk. Actual production displacement of therapist dosage verification tasks is minimal despite decades of planning systems.
Sector adoption velocityclaude-sonnet-53/5Healthcare/radiation oncology has adopted computerized planning and verification systems widely, but adoption of fully autonomous AI dose calculation without human check remains rare in clinical production settings.
Augmentation potentialclaude-haiku-4-5-202510014/5Modern treatment planning and dose calculation software substantially augments therapist productivity by automating complex physics computations, flagging dose anomalies, and streamlining plan review. The human therapist remains in the loop for final verification and clinical judgment, making this a strong augmentation scenario.
Augmentation potentialclaude-sonnet-54/5AI-assisted dosimetry and automated record-and-verify systems meaningfully speed up and improve accuracy of dose calculations while therapists remain responsible for verification and delivery oversight.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist with dose calculations using treatment planning algorithms and patient data, the task requires real-time verification, equipment-specific adjustments, and safety sign-off that currently demands human oversight. End-to-end automation with ≥50% time savings at equal quality is not reliably achievable today without human validation at critical checkpoints.
Task automatabilityclaude-sonnet-52/5Dose calculation involves precise, protocol-driven arithmetic that software can support, but verifying actual delivered dose per session requires machine-specific readouts, patient setup verification, and clinical judgment that current AI cannot fully own end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy is heavily regulated by medical boards and radiation safety authorities; dosage delivery must be calculated, verified, and signed off by licensed radiation therapists or medical physicists. Legal liability for treatment errors is asymmetric and severe, creating a hard barrier to unsupervised AI automation.
Adoption barriersclaude-sonnet-55/5Radiation dose calculation and delivery verification is heavily regulated (FDA, state licensure, medical physics oversight) and errors carry severe patient safety and liability consequences, requiring licensed professional sign-off.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based dose calculation systems require significant upfront investment (treatment planning software licenses, integration, training) and ongoing oversight costs. The marginal cost per session remains comparable to or higher than the direct labor cost of a radiation therapist performing the calculation, especially accounting for validation.
Cost vs. human wageclaude-sonnet-52/5Dosimetry software is already embedded in linear accelerator systems as sunk infrastructure cost, but the human calculation/verification role adds relatively modest incremental cost, so AI does not clearly beat human cost by an order of magnitude for this narrow sub-task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Treatment planning systems exist and can compute dosages, but deployed products typically function as decision-support tools requiring therapist verification rather than autonomous dose calculators. Clinical workflows mandate human review and authorization before dose delivery, reflecting regulatory and liability constraints rather than technical maturity.
Technical feasibility todayclaude-sonnet-53/5Treatment planning systems and record-and-verify software already compute and cross-check dosages in production, but these are calculation-support tools requiring therapist verification rather than fully autonomous dose determination.

Photograph treated area of patient and process film.

21

CI 16–25 · exposure 20 · augmentation 38 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Radiation therapy departments are specialized, regulated environments with slower digital transformation than general IT sectors. Adoption of automated imaging verification remains limited to large academic medical centers; most facilities continue manual processes due to cost and regulatory caution.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging workflows are adopting AI for image analysis and processing, but the physical photographing/positioning step in radiation therapy remains largely manual with slow adoption of full automation.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted image analysis (e.g., automated anatomical landmark detection, positioning feedback) can help technicians verify setup more quickly, but the human therapist remains responsible for final clinical judgment and regulatory sign-off on patient positioning.
Augmentation potentialclaude-sonnet-52/5AI can assist with digital image processing, storage, and quality checks after acquisition, but offers little help with the physical photographing step itself.
Task automatabilityclaude-haiku-4-5-202510012/5Photographing and basic film processing are mechanically automatable, but radiation therapy requires precise positioning tied to clinical protocols and patient-specific anatomy. Current AI can assist with image capture but cannot fully replace the clinical judgment of positioning and the regulatory need for human verification of treatment areas.
Task automatabilityclaude-sonnet-52/5Physically positioning a patient and operating imaging/photographic equipment plus film processing requires hands-on physical presence that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Radiation therapy is heavily regulated (FDA, state licensing boards, accreditation standards). A licensed radiation therapist must verify treatment setup and patient positioning for legal and safety liability; autonomous systems cannot substitute without explicit regulatory approval and physician oversight.
Adoption barriersclaude-sonnet-54/5This involves direct patient contact, positioning, and clinical documentation tied to licensed radiation therapy practice, creating strong regulatory and safety barriers to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5Automated imaging and film processing systems (when deployed) are capital-intensive and require integration into clinic workflows. The cost of hardware, maintenance, and integration often exceeds or approaches the cost of a technician performing these tasks, especially at smaller facilities.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for the physical act of imaging and processing, so the human cost is the only real option, making AI comparatively non-competitive here.
Technical feasibility todayclaude-haiku-4-5-202510012/5While automated imaging systems exist in some radiation therapy departments, they lack reliable end-to-end deployment for consistent clinical use. Current AI image processing products cannot independently verify correct anatomical positioning or validate treatment readiness without human oversight.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product autonomously photographs patients and processes film in clinical radiation therapy settings; this remains a manual technical task.

Enter data into computer and set controls to operate or adjust equipment or regulate dosage.

20

CI 20–20 · exposure 25 · augmentation 63 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare organizations adopt AI cautiously in high-stakes clinical domains; while data management tools see uptake, autonomous control of radiation delivery remains pilot-stage due to regulatory, liability, and patient safety concerns.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially radiation oncology, is a slow-adopting, highly regulated sector where AI is used for planning support but not for direct equipment operation or dosage control in production.
Augmentation potentialclaude-haiku-4-5-202510014/5AI can assist substantially by automating data entry, flagging protocol deviations, and suggesting dosage calculations based on imaging and patient records, materially speeding the therapist's workflow while the human retains control over final adjustments and safety decisions.
Augmentation potentialclaude-sonnet-53/5AI-assisted treatment planning and automated data entry/verification systems help therapists reduce errors and speed workflow, though the core control-setting task still requires direct human execution.
Task automatabilityclaude-haiku-4-5-202510012/5Data entry into computer systems is automatable, but setting controls and regulating dosage require real-time clinical judgment tied to patient anatomy and treatment protocols that vary case-by-case. Current systems cannot reliably perform the full end-to-end task without extensive human oversight.
Task automatabilityclaude-sonnet-52/5Setting treatment machine controls and dosage parameters requires precise verification against physician-prescribed plans and physical patient setup, which current AI cannot safely execute end-to-end without human operation.atorship.rationale.a rationale.a rationale.a
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy is heavily regulated by medical physics oversight, state licensure, and FDA/medical device approvals; a licensed radiation therapist must typically sign off on or directly execute dosage control per clinical protocols, creating hard legal and safety barriers.
Adoption barriersclaude-sonnet-55/5Radiation therapy delivery is heavily regulated, requiring licensed radiation therapists to operate equipment and verify dosage, with severe liability for errors, making full automation legally and practically barred.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration costs, ongoing oversight, and the need for human verification offset savings from automating data entry alone. Full automation is not yet cost-competitive with a trained radiation therapist performing the complete task.
Cost vs. human wageclaude-sonnet-52/5Linear accelerators and dosage control systems require certified human operators for safety and regulatory compliance, so AI cannot yet substitute cheaply for this labor without added oversight costs.
Technical feasibility todayclaude-haiku-4-5-202510012/5While data entry automation exists, integrated systems that autonomously adjust radiation equipment dosage in clinical practice do not reliably operate without human validation; regulatory approval and liability concerns limit deployment of fully automated dosage regulation.
Technical feasibility todayclaude-sonnet-52/5Treatment planning software and record-and-verify systems assist with dose calculation and data transfer, but actual equipment operation and control-setting remain manual actions performed by licensed therapists in production clinics.

Check for side effects, such as skin irritation, nausea, or hair loss to assess patients' reaction to treatment.

14

CI 3–25 · exposure 13 · augmentation 50 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare automation adoption in oncology is moderate; while supportive tools are piloted, direct clinical assessment by licensed staff remains entrenched due to regulatory, liability, and quality-of-care norms, limiting rapid replacement.
Sector adoption velocityclaude-sonnet-52/5Healthcare delivery, especially hands-on clinical assessment in oncology, adopts AI slowly due to regulatory, safety, and liability constraints despite digitization in imaging/diagnostics.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could meaningfully assist by flagging visual skin changes from photos, summarizing patient-reported symptoms from questionnaires, or highlighting trends in prior visits, allowing the therapist to focus deeper assessment on high-risk areas.
Augmentation potentialclaude-sonnet-53/5AI can help by flagging symptom patterns from patient-reported data, documentation assistance, or decision support, but the core physical assessment still needs direct human observation.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can assist in analyzing documented symptoms and patterns from medical records, the task requires direct physical examination (assessing skin condition, observing patient presentation) and nuanced clinical judgment about treatment tolerance that current AI systems cannot perform end-to-end without substantial human involvement.
Task automatabilityclaude-sonnet-51/5This requires direct physical examination, palpation, and patient interview to assess symptoms and severity, which current AI cannot perform end-to-end without a human present.
Adoption barriersclaude-haiku-4-5-202510014/5Radiation therapists are licensed professionals, and assessing treatment side effects is part of clinical care; liability and regulatory requirements (FDA oversight of diagnostic aids, clinical responsibility) create substantial barriers to full automation without licensed professional sign-off.
Adoption barriersclaude-sonnet-55/5Clinical assessment of patient reactions to radiation treatment requires licensed medical personnel with legal and safety accountability, an absolute barrier to automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of AI-assisted screening tools, integration, and mandatory clinician review would be comparable to or exceed the cost of a radiation therapist performing direct assessment, especially given the need for human verification of any automated flagging.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical assessment, so cost comparison favors the human entirely; AI cannot replace the labor being measured.
Technical feasibility todayclaude-haiku-4-5-202510012/5Computer vision systems can detect some visual skin changes from images, and NLP can extract symptom reports from notes, but no deployed clinical product reliably performs the full assessment (physical exam + integrated judgment) as a standalone system; human oversight remains mandatory.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously conducts physical side-effect assessments of radiation therapy patients; this remains a clinical, hands-on task.

Check radiation therapy equipment to ensure proper operation.

13

CI 0–25 · exposure 13 · augmentation 38 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Radiation therapy is a heavily regulated clinical domain with limited digitization and slow technology adoption. Equipment manufacturers provide some digital tools, but uptake of autonomous checking systems remains minimal due to regulatory and safety-critical constraints.
Sector adoption velocityclaude-sonnet-52/5Healthcare, particularly radiation oncology, has slow AI adoption for physical/safety-critical tasks due to regulatory requirements and the high stakes of equipment malfunction.
Augmentation potentialclaude-haiku-4-5-202510013/5AI systems can assist by automated data logging, trend analysis, and anomaly alerting that helps technicians focus investigations, reducing time spent on routine checks. However, the human must remain in the loop for final certification and decision-making.
Augmentation potentialclaude-sonnet-52/5AI-driven sensors and software can flag anomalies or log diagnostic data to assist technologists, but the core physical inspection and judgment calls remain human-driven with limited AI augmentation currently deployed.
Task automatabilityclaude-haiku-4-5-202510012/5While AI systems can monitor sensor data and flag anomalies in equipment logs, the task requires hands-on physical inspection, calibration verification, and judgment calls about whether equipment meets safety thresholds—actions that demand human intervention on-site. Current AI cannot perform these checks autonomously.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical safety check involving mechanical inspection, calibration verification, and machine warm-up procedures that require physical presence and manipulation of equipment; current AI cannot perform physical inspection tasks.
Adoption barriersclaude-haiku-4-5-202510014/5Radiation therapy equipment checks are governed by strict regulatory standards (NRC, state physics boards) and accreditation bodies (AAPM, ASTRO) that require a licensed Medical Physicist or Radiation Therapist to sign off on safety. Liability and patient safety concerns create hard legal and institutional barriers to full automation.
Adoption barriersclaude-sonnet-55/5Radiation therapy equipment checks are subject to strict regulatory, licensing, and safety protocols requiring certified radiation therapists or medical physicists to verify and sign off on equipment status before patient treatment.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI monitoring systems require significant capital investment and integration into clinical workflows, plus continuous human oversight. The loaded cost of a radiation therapist's time to do a full equipment check is likely lower than deploying and maintaining a comprehensive automated system.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical verification task, so cost comparison favors the human by default since AI cannot yet execute the task at all.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some hospitals use automated monitoring systems and alerts for radiation equipment, but these are narrow in scope (specific parameters only) and require human technicians to investigate and validate findings. No deployed AI system independently certifies equipment readiness.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical equipment safety checks on linear accelerators or radiation therapy machines; this remains a manual, regulated procedure performed by trained technologists.

Educate, prepare, and reassure patients and their families by answering questions, providing physical assistance, and reinforcing physicians' advice regarding treatment reactions or post-treatment care.

13

CI 0–25 · exposure 13 · augmentation 38 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of AI for high-touch patient education and reassurance remains minimal; institutions continue to rely on licensed clinicians for these functions, and there is strong professional and regulatory resistance to substitution.
Sector adoption velocityclaude-sonnet-52/5Healthcare delivery, especially in oncology treatment settings, has been slow to adopt AI for direct patient interaction due to regulatory, safety, and trust concerns, though administrative uses are growing.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with pre-prepared educational materials or reminders about post-treatment care, but the core reassurance and adaptive counseling functions require human judgment and presence; augmentation potential is limited.
Augmentation potentialclaude-sonnet-53/5AI can provide useful support such as generating easy-to-understand educational materials, answering common patient questions, and summarizing information for family members, improving efficiency in the education component.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires complex interpersonal communication, empathy, physical assistance, and the ability to address patient concerns in real-time. Current AI systems cannot reliably handle the nuanced, context-dependent reassurance and counseling needed, nor can they provide physical assistance to patients.
Task automatabilityclaude-sonnet-52/5This task blends emotional reassurance, physical assistance, and personalized patient education that requires human presence and empathy; AI can support information delivery but cannot replace the interpersonal and physical components.
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapists are licensed healthcare professionals, and patient education and reassurance during sensitive medical treatment is legally and ethically bound to human practitioners. Regulatory frameworks and standard of care requirements create hard barriers to automation of this task.
Adoption barriersclaude-sonnet-54/5Direct patient care, physical assistance, and reinforcing physician-directed treatment advice typically require a licensed radiation therapist, with liability and clinical protocol considerations limiting substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of deploying AI to simulate human reassurance and physical assistance—including oversight, liability management, and integration into clinical workflows—far exceeds the cost of a radiation therapist providing this care directly.
Cost vs. human wageclaude-sonnet-52/5While AI-driven FAQ/chat tools are cheap, the physical assistance and in-person reassurance components still require a human therapist, so cost savings are limited to a small slice of the task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs this task end-to-end in clinical settings. While chatbots exist for basic health Q&A, they lack the medical credibility, contextual understanding, and human presence required to educate and reassure radiation therapy patients about their specific treatment.
Technical feasibility todayclaude-sonnet-52/5Chatbots and patient education tools exist for general information, but no deployed product reliably handles reassurance, physical assistance, or nuanced clinical Q&A in radiation therapy settings.

Provide assistance to other healthcare personnel during dosimetry procedures and tumor localization.

13

CI 0–25 · exposure 13 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Radiation therapy is a highly specialized, regulated medical domain with low digitization of core clinical tasks and strong reliance on licensed professionals. Adoption of AI agents into patient-facing oncology workflows has been negligible despite pressure to improve efficiency.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially radiation oncology, has slower AI adoption for clinical procedural tasks due to regulatory and safety constraints, though AI-assisted contouring is gaining some traction in planning workflows.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist with pre-procedure planning (image analysis, documentation) or post-procedure record-keeping, but during active dosimetry and tumor localization, the human therapist's judgment, spatial awareness, and safety vigilance are irreplaceable. Assistance is limited to peripheral tasks.
Augmentation potentialclaude-sonnet-54/5AI-based auto-segmentation and treatment planning tools meaningfully speed up the dosimetry and localization process, augmenting the work of therapists and dosimetrists even though humans remain central to execution.
Task automatabilityclaude-haiku-4-5-202510011/5Assisting healthcare personnel during dosimetry and tumor localization requires physical presence, real-time coordination, contextual judgment about patient positioning, and adaptive response to clinical situations—none of which current AI can perform end-to-end without human supervision. The task is fundamentally collaborative and embodied.
Task automatabilityclaude-sonnet-52/5This is a hands-on assistive clinical task involving patient positioning, imaging setup, and coordination with dosimetrists/physicians; AI can support parts (image analysis, contouring) but cannot perform the physical assistance and interpersonal coordination end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Radiation oncology is heavily regulated (state licensure, federal oversight via FDA/NRC, hospital credentialing), patients require direct human contact and monitoring, and liability for dosimetry errors is severe. A licensed healthcare worker must legally participate in and oversee these procedures.
Adoption barriersclaude-sonnet-54/5Radiation therapy involves licensed personnel, strict regulatory oversight (radiation safety), and high liability for errors, requiring certified therapists to be physically present and accountable.
Cost vs. human wageclaude-haiku-4-5-202510011/5The loaded wage for a radiation therapy technician is substantial, and the overhead of integrating AI agents into operating clinical workflows—including validation, liability, and human oversight—would exceed the savings from partial automation if any were possible.
Cost vs. human wageclaude-sonnet-52/5AI contouring/planning tools reduce some labor costs, but the physical presence and coordination required still necessitate human staff, so overall cost savings are limited relative to full task substitution.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously assist medical personnel in radiation oncology procedures. This task requires embodied presence, safety-critical decision-making, and real-time interaction with equipment and patients that current technology cannot reliably execute in production clinical settings.
Technical feasibility todayclaude-sonnet-52/5AI tools exist for auto-contouring and treatment planning support, but no deployed product performs the physical/procedural assistance role of a radiation therapist during localization and dosimetry in production.

Prepare or construct equipment, such as immobilization, treatment, or protection devices.

11

CI 5–16 · exposure 8 · augmentation 38 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Radiation therapy departments are conservative adopters for safety-critical tasks; automation of device fabrication remains minimal and largely experimental. Adoption is slow outside research settings given regulatory oversight and patient safety requirements.
Sector adoption velocityclaude-sonnet-51/5Healthcare's physical, hands-on clinical tasks like device fabrication show minimal AI adoption; this is a low-digitization, high-touch task.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted design tools (generative CAD, simulation for immobilization geometry) can augment therapist productivity in planning and initial device design, though physical construction and final assembly remain human-dependent.
Augmentation potentialclaude-sonnet-52/5AI can assist with design optimization or 3D-printing templates for immobilization devices, but the physical construction and fitting remain manual with limited AI augmentation currently.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist in design optimization and material selection, the physical construction and precise customization of immobilization and protection devices requires hands-on manufacturing, fitting, and quality assurance that current AI cannot perform end-to-end. The task involves tangible, patient-specific fabrication rather than information processing.
Task automatabilityclaude-sonnet-51/5Physically fabricating and fitting immobilization masks, molds, or shielding blocks requires manual craftsmanship and direct patient contact that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510014/5Equipment used in radiation therapy is subject to FDA regulation and quality assurance standards; medical device manufacturing and quality sign-off typically require a licensed professional or engineer. Liability for equipment failure in patient treatment creates high barriers to autonomous AI construction.
Adoption barriersclaude-sonnet-54/5Radiation therapy devices affecting patient safety and dose delivery require licensed therapist involvement and quality assurance sign-off, creating strong regulatory and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems would require significant overhead (robotics, specialized manufacturing integration, human oversight) that currently exceeds the cost of a trained radiation therapist or technician performing direct fabrication and assembly.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical fabrication task, so cost comparison favors the human therapist by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably manufactures or constructs medical radiation devices independently. While CAD/CAM tools exist, they are not autonomous—they remain human-directed design and fabrication aids, not end-to-end construction systems.
Technical feasibility todayclaude-sonnet-51/5No deployed AI or robotic product constructs or fits patient-specific immobilization or shielding devices in clinical practice today; this remains a manual therapist task.

Position patients for treatment with accuracy, according to prescription.

6

CI 0–11 · exposure 8 · augmentation 50 · importance 4.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare, especially oncology departments, has slow technology adoption; regulatory requirements, safety liability, and the need for human clinical judgment mean patient positioning remains almost entirely manual despite decades of robotics development.
Sector adoption velocityclaude-sonnet-52/5Healthcare delivery, especially hands-on clinical procedures, adopts automation slowly due to safety, regulatory, and physical execution constraints, though imaging/planning software adoption is growing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI-assisted guidance through real-time imaging overlay, auto-alignment feedback, or planning tools could help therapists position patients more consistently and efficiently, though the core task remains human-performed.
Augmentation potentialclaude-sonnet-53/5AI-assisted imaging and positioning verification systems (e.g., surface-guided radiotherapy) help therapists confirm and fine-tune positioning accuracy, improving precision while the human still performs the physical task.
Task automatabilityclaude-haiku-4-5-202510012/5Positioning patients requires physical manipulation and precise spatial alignment verified through imaging feedback. Current AI systems cannot independently handle the embodied coordination, patient comfort assessment, and real-time adjustment needed, though AI could assist with guidance or planning.
Task automatabilityclaude-sonnet-51/5Physical positioning of patients on a treatment table requires manual handling, verification against immobilization devices, and real-time adjustment that current AI cannot execute end-to-end without robotics far beyond deployed systems.'
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy is heavily regulated; positioning patients is a licensed clinical task that must be performed or directly supervised by a qualified radiation therapist under medical board rules. Legal liability and patient safety requirements create hard barriers to full automation.
Adoption barriersclaude-sonnet-55/5Patient positioning for radiation treatment is a licensed clinical task with direct patient contact, safety-critical accuracy requirements, and regulatory/liability constraints requiring a certified radiation therapist.
Cost vs. human wageclaude-haiku-4-5-202510011/5Deploying automated patient positioning would require expensive robotic hardware, integration with treatment systems, safety validation, and continuous human oversight—far exceeding the cost of a trained therapist performing the task directly.
Cost vs. human wageclaude-sonnet-51/5There is no AI system substituting for the physical labor and clinical judgment involved, so cost comparison favors the human therapist entirely at this time.
Technical feasibility todayclaude-haiku-4-5-202510011/5No production system today autonomously positions patients for radiation therapy. This task requires licensed medical personnel oversight and direct patient contact; deployed robotic positioning systems exist only as narrow, supervised aids in research or limited clinical settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously positions patients for radiation therapy; imaging-based positioning verification tools exist but require a human therapist to physically place and adjust the patient.

Act as liaison with physicist and supportive care personnel.

4

CI 0–7 · exposure 0 · augmentation 25 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare sectors are slower to automate soft coordination tasks due to regulatory, liability, and safety constraints. Liaison roles remain predominantly human-staffed even in digitally advanced hospital systems.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially clinical radiation oncology, has slower AI adoption for interpersonal/coordination roles compared to purely administrative or diagnostic tasks, with pilots focused on imaging/planning rather than staff liaison work.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist by scheduling coordination meetings, summarizing clinical notes, or flagging communication gaps, but the core liaison function—negotiating priorities and building consensus—remains inherently human and requires minimal AI assistance to perform effectively.
Augmentation potentialclaude-sonnet-52/5AI tools like shared documentation systems, scheduling aids, or communication platforms can support information flow, but they don't meaningfully transform the interpersonal liaison function itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task is fundamentally a human liaison role requiring judgment, relationship-building, and context-dependent communication between specialized professionals. Current AI cannot autonomously mediate complex interpersonal coordination, resolve conflicts, or make discretionary decisions about resource allocation across clinical teams.
Task automatabilityclaude-sonnet-51/5This is an interpersonal coordination and communication task requiring real-time clinical judgment, relationship management, and physical presence in a care team context, which current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Medical radiation therapy is heavily regulated; coordination between physicists and care teams typically requires documented human accountability, licensure, and professional judgment. Legal and regulatory frameworks mandate human responsibility for patient safety decisions.
Adoption barriersclaude-sonnet-54/5Radiation therapy involves licensed personnel, strict safety protocols, and multidisciplinary accountability, meaning liaison communications tied to patient care decisions require qualified human staff and carry liability implications.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of developing and maintaining AI systems capable of meaningful liaison work, plus the human oversight required, far exceeds the wage of a liaison coordinator. The error costs (miscommunication in radiation therapy) would be catastrophic.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this liaison role, so no meaningful cost comparison exists; the human is the only viable option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs autonomous liaison duties in clinical settings. Such work requires genuine understanding of organizational dynamics, clinical priorities, and professional relationships that current systems cannot reliably replicate.
Technical feasibility todayclaude-sonnet-51/5No deployed product acts as an autonomous liaison between clinical staff members in oncology care settings; this remains firmly a human coordination function.

Train or supervise student or subordinate radiotherapy technologists.

4

CI 0–7 · exposure 0 · augmentation 38 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare, especially radiotherapy, is characterized by low automation adoption for human-facing supervisory and training roles. Regulatory requirements, patient-safety mandates, and the critical nature of clinical education mean this sector lags in automating training functions.
Sector adoption velocityclaude-sonnet-52/5Healthcare clinical training remains a slow-adopting, highly regulated, human-centric domain despite AI's broader gains in some administrative healthcare functions.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with scheduling, documentation, or providing reference materials during training, but adds marginal value to the core act of teaching competency and evaluating technologist performance in a safety-critical clinical setting.
Augmentation potentialclaude-sonnet-53/5AI-based simulation tools, e-learning modules, and knowledge assessment platforms can meaningfully support training content delivery and skill assessment, augmenting but not replacing human supervisors.
Task automatabilityclaude-haiku-4-5-202510011/5Training and supervision require real-time judgment, interpersonal correction, and adaptive mentoring based on individual learner needs—capabilities beyond current AI systems. AI cannot reliably assess competence, provide nuanced feedback, or build the trust relationships essential to effective training.
Task automatabilityclaude-sonnet-51/5Training and supervising human trainees involves live demonstration, hands-on correction, mentorship, and real-time clinical judgment that current AI cannot perform end-to-end.'
Adoption barriersclaude-haiku-4-5-202510015/5Radiotherapy is a highly regulated clinical domain; only licensed radiation therapists can legally supervise and train technologists. Liability, patient safety, and professional licensing requirements create hard legal barriers preventing AI substitution or independent operation.
Adoption barriersclaude-sonnet-54/5Clinical training and supervision typically require credentialed, licensed professionals accountable for trainee competency and patient safety, creating strong regulatory and liability barriers.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI systems cannot perform this task end-to-end, so cost comparison is moot. A qualified radiation therapist trainer commands professional wages that AI cannot displace because AI cannot execute training and supervision at all today.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this supervisory/teaching role, so cost comparison favors the human by default.
Technical feasibility todayclaude-haiku-4-5-202510011/5No production system today can autonomously train or supervise radiotherapy technologists. This task requires human oversight of safety-critical clinical decisions and relies on licensure, professional accountability, and legal responsibility that cannot be automated.
Technical feasibility todayclaude-sonnet-51/5No deployed product supervises or trains clinical radiotherapy staff autonomously; at most AI provides supplementary training materials or simulations.

Follow principles of radiation protection for patient, self, and others.

3

CI 3–3 · exposure 0 · augmentation 50 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI is moderate and cautious; radiation therapy remains a highly supervised, credentialed field where regulatory and liability constraints slow any shift from human-centric workflows. Pilot projects exist but production displacement is minimal.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging/radiotherapy adopts AI for planning and image analysis, but physical safety protocol execution sees minimal AI-driven displacement.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist by providing real-time dose calculations, flagging deviations from treatment plans, and recommending shielding adjustments, but the therapist must interpret and act on these suggestions. Augmentation is meaningful but secondary to human decision-making.
Augmentation potentialclaude-sonnet-53/5AI can support dose calculation, treatment planning, and monitoring systems that help therapists apply protection principles more precisely, though the core protective actions remain human-performed.
Task automatabilityclaude-haiku-4-5-202510011/5Following radiation protection principles requires real-time judgment about dose levels, patient positioning, shielding placement, and environmental hazards—domain-specific decisions that demand continuous human oversight. Current AI cannot autonomously ensure compliance with physics principles and safety protocols in a clinical setting where errors have immediate health consequences.
Task automatabilityclaude-sonnet-51/5This task requires physical presence, equipment operation, shielding decisions, and real-time judgment about radiation safety that cannot be executed end-to-end by current AI systems.apoi
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy is heavily regulated by the NRC, state health boards, and OSHA; a licensed radiation therapist must legally perform or directly oversee dose delivery and safety protocols. Liability asymmetry is extreme—AI errors in radiation safety cause irreversible harm, creating a hard barrier to automation.
Adoption barriersclaude-sonnet-55/5Radiation safety is heavily regulated, requires licensed practitioners, and involves direct patient safety and legal liability, making human execution mandatory.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of developing, validating, and maintaining an AI system for radiation safety, combined with legal liability and regulatory certification costs, far exceeds the loaded wage of a radiation therapist who performs this task as part of standard practice.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical safety task, so no cost comparison favors AI; the human must be present regardless.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product independently manages radiation protection compliance in therapy settings. While dose calculation software exists, the task of *following* protection principles—monitoring, adjusting, enforcing safety protocols—remains human-dependent with no production system replacing this judgment.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical radiation protection practices; this remains a hands-on clinical safety procedure performed by trained staff.

Conduct most treatment sessions independently, in accordance with the long-term treatment plan and under the general direction of the patient's physician.

1

CI 0–3 · exposure 0 · augmentation 50 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI is cautious and heavily regulated. While AI tools support planning and diagnosis, treatment delivery remains stubbornly human-centric due to regulatory and liability constraints; production-scale autonomous treatment sessions do not exist.
Sector adoption velocityclaude-sonnet-51/5Healthcare delivery involving direct physical patient care is a slow-adopting sector for full automation, with AI confined to planning and image analysis rather than treatment administration.
Augmentation potentialclaude-haiku-4-5-202510013/5AI assists radiation therapists through image guidance, treatment plan optimization, and positioning feedback, raising the quality and precision of human-delivered treatment. However, augmentation is limited to planning and decision support, not the delivery itself.
Augmentation potentialclaude-sonnet-53/5AI can assist with treatment planning, image registration, and quality assurance checks that support the therapist, but the actual conduct of the session remains manually performed.
Task automatabilityclaude-haiku-4-5-202510011/5Radiation therapy treatment sessions require real-time clinical judgment, patient interaction, positioning, safety monitoring, and immediate response to complications—tasks that demand human presence and accountability. No current AI system can conduct these sessions end-to-end independently.
Task automatabilityclaude-sonnet-51/5This task requires physically positioning patients, operating radiation equipment, monitoring patient condition in real time, and responding to physical/medical needs during treatment—none of which current AI systems can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy is heavily regulated; only licensed radiation therapists can legally conduct treatment sessions. Liability, patient safety requirements, and direct regulatory mandates create hard barriers that prevent full substitution.
Adoption barriersclaude-sonnet-55/5Radiation therapy delivery is heavily regulated, requires licensed therapists, direct physician oversight, and involves high liability from radiation exposure errors, making legal and safety barriers extremely high.
Cost vs. human wageclaude-haiku-4-5-202510011/5Radiation therapy equipment, licensing, liability, and the irreplaceable human oversight make automation economically infeasible. The loaded cost of a licensed therapist conducting treatment is far lower than any conceivable AI alternative that maintains safety and legal compliance.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical, hands-on clinical task, so no meaningful cost comparison exists—human labor is the only option.
Technical feasibility todayclaude-haiku-4-5-202510011/5While AI assists in treatment planning and image analysis, no deployed system currently conducts actual treatment sessions. The task requires licensed human radiation therapists to physically deliver and oversee treatment in real-world clinical practice.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently conducts radiation therapy sessions; AI is used only for planning/imaging support, not hands-on treatment delivery.

Implement appropriate follow-up care plans.

1

CI 0–3 · exposure 0 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare automation lags even information-intensive sectors; radiation therapy is highly regulated and clinically conservative. Adoption of AI for independent care planning decisions remains minimal in production settings, with most healthcare organizations still in early pilot phases.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially direct patient care in oncology, adopts AI slowly due to regulatory, safety, and workflow integration challenges, with most AI use confined to imaging or scheduling support.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully assist by retrieving evidence-based guidelines, flagging outlier patient metrics requiring review, or generating draft plan templates, which a therapist then refines. This augmentation is real but modest—the human therapist retains full analytical and decision-making responsibility.
Augmentation potentialclaude-sonnet-53/5AI can help draft follow-up schedules, flag risk factors from records, and summarize patient history, aiding therapists though not replacing their judgment.
Task automatabilityclaude-haiku-4-5-202510011/5Follow-up care plans for radiation therapy require individualized clinical judgment, patient-specific factors (disease progression, side effects, comorbidities), and adaptive decision-making that current AI cannot reliably execute end-to-end. While AI might assist in data gathering, the synthesis into a coherent, personalized care plan remains firmly within human clinical expertise.
Task automatabilityclaude-sonnet-51/5Implementing follow-up care requires clinical judgment, patient interaction, monitoring for treatment complications, and coordination with oncology teams—none of which current AI can execute end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Clinical care plan implementation is legally and ethically bound to licensed healthcare providers. Radiation therapists must sign off on follow-up protocols, and malpractice liability, regulatory oversight (state licensure, hospital credentialing), and the human-contact requirement for patient-specific decisions create hard barriers to autonomous substitution.
Adoption barriersclaude-sonnet-55/5Follow-up care in radiation oncology requires licensed clinical staff to assess patient status, manage side effects, and make care decisions—strong regulatory and liability barriers prevent full automation.
Cost vs. human wageclaude-haiku-4-5-202510011/5The all-in cost of AI systems for clinical care planning (models, infrastructure, human oversight, liability) exceeds the cost of a trained radiation therapist performing this task, given the low automation feasibility and high oversight burden.
Cost vs. human wageclaude-sonnet-51/5Human clinical oversight, licensure, and liability requirements make AI substitution costlier or infeasible relative to a therapist's wage for this task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product reliably generates or implements follow-up care plans for radiation therapy patients independently. Existing AI tools may support evidence retrieval or monitoring alerts, but no production system performs this clinical task autonomously in healthcare settings.
Technical feasibility todayclaude-sonnet-51/5No deployed product independently implements follow-up care plans for radiation therapy patients; this remains a clinician-led activity with AI only in research or decision-support roles.

Administer prescribed doses of radiation to specific body parts, using radiation therapy equipment according to established practices and standards.

0

CI 0–0 · exposure 0 · augmentation 38 · importance 4.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a laggard sector for clinical task automation due to regulatory complexity, liability concerns, and the requirement for licensed practitioners. Radiation therapy specifically involves controlled substances and high-consequence clinical decisions where automation adoption is nearly nonexistent.
Sector adoption velocityclaude-sonnet-51/5Healthcare delivery of physical treatments is a slow-adopting sector for full automation due to safety, regulatory, and physical infrastructure constraints.
Augmentation potentialclaude-haiku-4-5-202510012/5Current AI offers minimal augmentation for the core administration task itself. While AI may assist with treatment planning or image analysis upstream, it does not meaningfully improve a therapist's productivity during the actual radiation dose administration—the human remains fully responsible for execution and safety decisions.
Augmentation potentialclaude-sonnet-53/5AI assists with treatment planning, dose calculation, and image-guided targeting that support the therapist, but the hands-on administration itself sees limited direct AI augmentation.
Task automatabilityclaude-haiku-4-5-202510011/5Administering radiation therapy requires precise physical positioning of equipment, real-time patient monitoring, and immediate safety decisions that current AI cannot perform end-to-end. The task involves complex sensorimotor control and human judgment in a high-consequence clinical environment where no automation path achieves 50% time savings at equal quality.
Task automatabilityclaude-sonnet-51/5Physically administering radiation requires positioning patients, operating linear accelerators, and real-time safety monitoring in a physical clinical space, which current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Hard regulatory barriers prevent substitution: radiation therapy must be administered by a licensed Radiation Therapist under physician supervision, with legal and liability requirements that mandate human sign-off. Medical licensing, malpractice liability asymmetry, and FDA-regulated equipment create insurmountable legal barriers.
Adoption barriersclaude-sonnet-55/5Radiation delivery is tightly regulated, requires licensed radiation therapists, involves major liability/safety risk, and mandates direct human presence and oversight.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI cannot currently perform this task, making cost comparison irrelevant. Even partial automation would require expensive specialized hardware and regulatory validation, while the human labor cost remains fixed by licensing requirements and the clinical necessity of human oversight.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for the physical delivery of treatment, so no cost comparison favors AI; human labor plus equipment remains the only means of delivery.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs this task autonomously today. Radiation therapy equipment requires licensed human operators to physically position equipment, monitor patient comfort and safety, and make real-time clinical adjustments—functions that remain wholly human-dependent in current clinical practice.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously delivers radiation therapy to patients; this remains a hands-on clinical procedure performed by licensed therapists.

Observe and reassure patients during treatment and report unusual reactions to physician or turn equipment off if unexpected adverse reactions occur.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare facilities remain conservative in automation of patient-facing clinical tasks. Adoption of AI for observation and intervention during active treatment is negligible; human therapists remain standard practice across all surveyed settings.
Sector adoption velocityclaude-sonnet-51/5Healthcare, especially in-treatment bedside/procedural monitoring, is a slow-adopting, highly regulated, physically-grounded sector with minimal AI displacement of this specific function.
Augmentation potentialclaude-haiku-4-5-202510012/5Automated vital sign monitoring or alerts could assist a therapist by flagging anomalies, but the core task—reassurance, judgment, and intervention—is fundamentally human-centric. Augmentation potential is limited because the human therapist must remain present and in control.
Augmentation potentialclaude-sonnet-52/5Sensors and monitoring systems can alert therapists to physiological anomalies, offering modest assistance, but the core reassurance and judgment-based intervention remains unaided by AI.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires real-time human judgment, empathy, and physical intervention (turning off equipment) in response to patient distress. Current AI cannot reliably detect subtle patient distress signals, provide genuine reassurance, or make safety-critical decisions to halt treatment without human oversight.
Task automatabilityclaude-sonnet-51/5This requires real-time physical presence, emotional reassurance, and immediate safety intervention on a live patient undergoing radiation treatment—no current AI system can perform this end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy is heavily regulated; a licensed radiation therapist or physician must legally supervise and make treatment decisions. Patient safety, liability, and regulatory requirements (e.g., state licensure boards) create hard barriers to autonomous AI substitution.
Adoption barriersclaude-sonnet-55/5Direct patient safety monitoring during radiation exposure with authority to halt treatment is a licensed clinical responsibility with strict regulatory and liability requirements for human oversight.
Cost vs. human wageclaude-haiku-4-5-202510011/5The setup and liability costs of an AI monitoring system would far exceed the cost of a radiation therapist's salary, given the safety-critical nature of the task and the need for human oversight anyway. The all-in cost (hardware, integration, validation, insurance) strongly favors retaining human observation.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this function, so cost comparison favors the human by default; any AI monitoring add-on would be supplemental, not a replacement of comparable cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously observe patients during radiation therapy and make real-time clinical safety decisions. While computer vision could theoretically detect gross vital sign changes, no production system exists that integrates monitoring, interpretation, reassurance, and equipment control in a clinically validated manner.
Technical feasibility todayclaude-sonnet-51/5No deployed product observes patients physically, provides human reassurance, or physically halts radiation equipment in response to adverse reactions; this remains squarely a human safety role.

Store, sterilize, or prepare the special applicators containing the radioactive substance implanted by the physician.

0

CI 0–0 · exposure 0 · augmentation 13 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare, particularly nuclear medicine, lags in automation adoption for regulated radioactive handling; no sector trend shows meaningful displacement of this task through AI or robotics.
Sector adoption velocityclaude-sonnet-51/5Healthcare physical/procedural tasks involving radioactive materials show minimal AI adoption due to safety, regulatory, and physical manipulation requirements.
Augmentation potentialclaude-haiku-4-5-202510012/5Limited opportunity for AI assistance; computer vision or inventory tracking could marginally help with storage organization or applicator tracking, but the core preparation and sterilization work remains manual and human-centered.
Augmentation potentialclaude-sonnet-51/5AI offers essentially no meaningful assistance for the physical sterilization, storage, or preparation of radioactive applicators.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physically handling and preparing radioactive material within controlled clinical settings, demanding human dexterity, real-time safety compliance, and judgment about sterility and material integrity. Current AI systems cannot autonomously manipulate physical objects with radioactive substances or verify sterilization outcomes reliably.
Task automatabilityclaude-sonnet-51/5This is a hands-on physical task requiring precise handling of radioactive materials and sterile equipment; no current AI system can physically manipulate, sterilize, or store radioactive applicators.
Adoption barriersclaude-haiku-4-5-202510015/5Nuclear Regulatory Commission (NRC) and state radiation control regulations require licensed radiation therapists to handle and prepare radioactive materials; liability for mishandling radioactive substances creates hard legal and safety barriers to full automation.
Adoption barriersclaude-sonnet-55/5Handling radioactive substances requires licensed radiation therapists/technologists under strict regulatory (NRC, state) oversight, with legal requirements for certified personnel to perform such tasks.
Cost vs. human wageclaude-haiku-4-5-202510011/5Robotics capable of handling radioactive materials with required precision and sterile protocols are extremely expensive to acquire, validate, and maintain—far exceeding the cost of trained radiation therapist labor for this task.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute for this physical task, so AI cost is not comparable—human labor with specialized equipment is the only option today.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs autonomous handling, sterilization, or preparation of radioactive medical applicators in production. This remains a specialized clinical manual task with no commercial automation solutions.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product performs physical sterilization or handling of radioactive medical devices; this remains entirely a manual clinical procedure.

Assist in the preparation of sealed radioactive materials, such as cobalt, radium, cesium, or isotopes, for use in radiation treatments.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 3.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare AI adoption in radiation oncology focuses on planning and imaging analysis rather than physical material handling; the sealed-source preparation task has seen no meaningful AI displacement and operates in a conservative, highly regulated environment.
Sector adoption velocityclaude-sonnet-51/5Healthcare physical/clinical procedures involving radioactive material handling show minimal AI adoption due to safety, regulatory, and physical constraints.
Augmentation potentialclaude-haiku-4-5-202510012/5While AI might assist with documentation, inventory tracking, or protocol verification adjacent to this task, it offers minimal direct assistance with the core activity of physically preparing radioactive materials under strict safety protocols.
Augmentation potentialclaude-sonnet-52/5AI may assist with dose calculation, scheduling, or documentation support, but offers negligible assistance with the physical preparation and handling of radioactive materials itself.
Task automatabilityclaude-haiku-4-5-202510011/5This task involves hands-on physical manipulation of hazardous sealed radioactive materials in controlled clinical settings, requiring precise spatial coordination, physical handling of equipment, and real-time safety monitoring that current AI systems cannot perform end-to-end.
Task automatabilityclaude-sonnet-51/5This involves physical handling and preparation of hazardous radioactive materials requiring precise manual manipulation, safety protocols, and physical presence; no AI system today performs this physical task.
Adoption barriersclaude-haiku-4-5-202510015/5Radiation therapy preparation is heavily regulated by the Nuclear Regulatory Commission and state agencies; only licensed radiation therapists are legally permitted to handle sealed radioactive sources, creating hard legal and liability barriers to any form of automation.
Adoption barriersclaude-sonnet-55/5Handling radioactive materials is tightly regulated, requiring licensed personnel, radiation safety certification, and strict chain-of-custody and safety protocols that legally require qualified humans.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI has no cost advantage here because the task fundamentally requires human physical presence and cannot be automated, making any comparison meaningless—human labor is the only viable option.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute for physical material handling, so the human cost is the only viable option, making AI infeasible rather than cheaper.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product can autonomously prepare sealed radioactive materials; this remains exclusively a human-performed task due to the physical, safety-critical, and regulatory nature of the work.
Technical feasibility todayclaude-sonnet-51/5No deployed product handles or prepares sealed radioactive sources; this remains a manual clinical/physical task performed by trained personnel.

Related occupations — Healthcare Practitioners & Technical

How to read this

A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.

What would change this score

New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.