Magnetic Resonance Imaging Technologists

29-2035.00
Median wage $95,480/yr43,390 employed (US)Rank #567 of 923 scored · top 61% by substitution

Operate Magnetic Resonance Imaging (MRI) scanners. Monitor patient safety and comfort, and view images of area being scanned to ensure quality of pictures. May administer gadolinium contrast dosage intravenously. May interview patient, explain MRI procedures, and position patient on examining table. May enter into the computer data such as patient history, anatomical area to be scanned, orientation specified, and position of entry.

Sub-scores

0–100 · band = confidence interval from rater disagreement

Substitution24
Exposure23
Augmentation45

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

24 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

13%

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%24

panel mean rating 2.0/5 → substitution pressure 24/100

Technical feasibility todayw 20%23

panel mean rating 1.9/5 → substitution pressure 23/100

Cost vs. human wagew 15%24

panel mean rating 2.0/5 → substitution pressure 24/100

Adoption barriersw 20%inverted — strong barriers lower the score28

panel mean rating 3.9/5 (barrier strength) → substitution pressure 28/100

Sector adoption velocityw 10%22

panel mean rating 1.9/5 → substitution pressure 22/100

Task breakdown (24 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.

Create backup copies of images by transferring images from disk to storage media or workstation.

96

CI 92–100 · exposure 100 · augmentation 38 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Healthcare organizations have rapidly adopted automated backup and DICOM archive systems as standard infrastructure for decades; this is now baseline practice in radiology departments rather than novel or pilot-stage automation.
Sector adoption velocityclaude-sonnet-55/5Healthcare imaging IT infrastructure (PACS) has already deeply and widely adopted automated storage/backup workflows for decades.
Augmentation potentialclaude-haiku-4-5-202510012/5Automated backup offers minimal augmentation to a technologist; the task is purely logistical data transfer with no analytical or judgment component where AI assistance would enhance human productivity.
Augmentation potentialclaude-sonnet-53/5While largely automated already, technologists still benefit from monitoring/management dashboards that streamline oversight of backup processes.
Task automatabilityclaude-haiku-4-5-202510015/5File transfer and backup operations are completely automatable end-to-end with standard tools (automated backup software, cloud storage services, scheduled tasks). Current systems can perform this with far greater than 50% time savings and at equal or superior quality, requiring no human intervention after initial setup.
Task automatabilityclaude-sonnet-55/5File transfer and backup of medical images is a straightforward, well-defined digital data-handling task easily scripted or automated via PACS systems and DICOM protocols.
Adoption barriersclaude-haiku-4-5-202510012/5While IT security policies and HIPAA compliance oversight create some procedural requirements around medical image handling, there are no licensing or legal barriers preventing automated backup systems; organizations routinely deploy these with standard technical controls rather than human-in-the-loop approval.
Adoption barriersclaude-sonnet-51/5This is a purely administrative/IT task with no licensure, liability, or human-contact requirement standing in the way of automation.
Cost vs. human wageclaude-haiku-4-5-202510015/5Automated backup and storage systems cost orders of magnitude less per execution than human technologist labor; a technologist hourly rate vastly exceeds the marginal cost of an automated backup job running on existing infrastructure.
Cost vs. human wageclaude-sonnet-55/5Automated backup software costs a tiny fraction of a technologist's hourly wage to run compared to manual file transfer time.
Technical feasibility todayclaude-haiku-4-5-202510015/5Mature, production-ready products demonstrably perform this task reliably at scale across healthcare and enterprise environments (DICOM backup solutions, hospital archive systems, automated tape and cloud backup platforms are in widespread clinical use).
Technical feasibility todayclaude-sonnet-55/5PACS/VNA systems already automate image archiving, backup, and storage routing as standard production functionality in virtually all modern radiology departments.

Schedule appointments for research subjects or clinical patients.

86

CI 79–92 · exposure 87 · augmentation 75 · importance 3.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Healthcare and research institutions are adopting AI-driven scheduling rapidly as part of broader clinic management and patient portal modernization. Adoption is particularly deep in large hospital systems, research centers, and digitized workflows—the core sectors where MRI technologists work.
Sector adoption velocityclaude-sonnet-54/5Healthcare administrative functions, including scheduling, have seen rapid adoption of automated and AI-assisted tools, though clinical fields overall adopt slower than pure information sectors.
Augmentation potentialclaude-haiku-4-5-202510014/5AI scheduling assistants significantly augment technologist productivity by flagging conflicts, suggesting optimal slots, auto-sending reminders, and handling rescheduling requests, freeing the technologist to focus on patient care and imaging tasks rather than calendar management.
Augmentation potentialclaude-sonnet-54/5AI scheduling assistants significantly reduce administrative burden on staff by handling routine bookings, reminders, and rescheduling, freeing technologists for clinical work.
Task automatabilityclaude-haiku-4-5-202510015/5Scheduling appointments is a well-defined, highly structured task with clear inputs (patient/subject availability, clinician availability, appointment duration) and outputs (booked time slots). Modern AI systems and calendar APIs can fully automate this end-to-end, easily meeting the 50% time-saving threshold with equal or better quality and reliability than manual scheduling.
Task automatabilityclaude-sonnet-54/5Appointment scheduling is a well-structured task involving calendar management, patient availability, and resource coordination that off-the-shelf scheduling software and AI agents can handle end-to-end with minimal setup.
Adoption barriersclaude-haiku-4-5-202510012/5While there are minimal hard legal barriers to automating scheduling itself, some friction exists from patient/subject preferences for human contact, institutional preferences for dual verification, and integration complexity with existing clinic workflows. However, these do not prevent automation.
Adoption barriersclaude-sonnet-52/5No licensure is required for scheduling; the main friction is integration with EHR/RIS systems and occasional need for human judgment on complex or urgent cases.
Cost vs. human wageclaude-haiku-4-5-202510015/5The cost of AI-driven scheduling (via calendar APIs, scheduling software, or agent inference) is orders of magnitude cheaper than the labor cost of a human scheduler or technologist spending time on manual booking and rescheduling tasks.
Cost vs. human wageclaude-sonnet-55/5Automated scheduling software costs a small fraction of a technologist's or scheduler's hourly wage once implemented, offering large per-transaction cost savings.
Technical feasibility todayclaude-haiku-4-5-202510015/5Deployed products (calendar automation, clinic management systems with AI scheduling, and general-purpose intelligent agents) reliably perform appointment scheduling at scale in healthcare and research settings today. Maturity is evident across medical practices, research institutions, and commercial platforms.
Technical feasibility todayclaude-sonnet-54/5Automated scheduling systems (e.g., Epic, Cerner integrated schedulers, chatbot-based booking tools) are already deployed widely in healthcare settings for patient appointment scheduling, including imaging departments.

Develop or otherwise produce film records of magnetic resonance images.

72

CI 70–75 · exposure 75 · augmentation 75 · importance 3.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510014/5Major health systems and imaging centers have rapidly integrated AI reconstruction and automated processing into workflows over the past 3–5 years. Adoption is broad in well-digitized institutional settings, though slower in smaller facilities.
Sector adoption velocityclaude-sonnet-54/5Healthcare imaging has broadly adopted digital PACS and automated reconstruction/output over the past two decades, making this a mature, well-adopted automation within radiology departments.
Augmentation potentialclaude-haiku-4-5-202510015/5AI significantly enhances technologist productivity by automating reconstruction, filtering, and export while technologists focus on acquisition protocol optimization, patient safety, and quality assurance. This creates a strong augmentation dynamic.
Augmentation potentialclaude-sonnet-53/5AI and automated software assist technologists by streamlining image formatting and distribution, though the technologist still manages quality control and workflow integration.
Task automatabilityclaude-haiku-4-5-202510014/5Current AI systems can reliably generate, post-process, and export MRI image records with minimal human intervention. DICOM conversion, image reconstruction, formatting, and archival can be largely automated, though quality review and clinical validation typically still require technologist oversight.
Task automatabilityclaude-sonnet-54/5Producing film/digital records from MRI acquisitions is a largely software-driven, standardized workflow (image reconstruction, formatting, printing/exporting to PACS) that modern imaging systems already automate heavily.PACS and scanner software handle most of this without technologist intervention beyond setup.
Adoption barriersclaude-haiku-4-5-202510013/5Clinical oversight, QA, and radiologist sign-off remain legally and organizationally required; MRI technologists are responsible for protocol selection and troubleshooting, creating moderate friction. However, the image production itself is not exclusively licensed.
Adoption barriersclaude-sonnet-52/5This specific sub-task (producing film/digital output) is not itself subject to licensing requirements, though it occurs within a clinical workflow overseen by licensed technologists and radiologists.
Cost vs. human wageclaude-haiku-4-5-202510014/5AI-driven reconstruction and automated DICOM processing cost a fraction of technologist labor per study. Infrastructure and software licenses are amortized across high volumes, making the all-in cost substantially lower than equivalent human time.
Cost vs. human wageclaude-sonnet-54/5Once integrated, automated digital record production has near-zero marginal cost compared to technologist time spent on manual film production/output tasks.
Technical feasibility todayclaude-haiku-4-5-202510014/5Commercial PACS systems and AI-powered image reconstruction software (GE's LFIRE, Siemens' advanced reconstruction) are deployed in clinical settings and demonstrably perform image processing and film record generation at production scale with high reliability.
Technical feasibility todayclaude-sonnet-54/5Modern MRI scanners and PACS systems already automate image reconstruction, formatting, and digital film production reliably in clinical production environments, though physical film printing where still used requires some manual steps.

Conduct inventories to maintain stock of clinical supplies.

47

CI 40–55 · exposure 42 · augmentation 63 · importance 4.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510013/5Large healthcare systems are adopting automated inventory and RFID tracking, but smaller radiology clinics and MRI centers often rely on manual or semi-automated methods; adoption is uneven and moderate rather than rapid and deep.
Sector adoption velocityclaude-sonnet-52/5Healthcare administrative/logistics functions adopt digital tools slowly relative to information sectors, with many imaging departments still using manual or semi-manual inventory processes.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-driven inventory dashboards, predictive stock-out alerts, and automated barcode scanning substantially assist technologists by reducing manual counting burden and flagging shortages before they occur, allowing focus on supply optimization rather than tedious counting.
Augmentation potentialclaude-sonnet-53/5Inventory management software and predictive reordering tools can meaningfully assist technologists in tracking usage patterns and flagging low stock, improving efficiency without full automation.
Task automatabilityclaude-haiku-4-5-202510012/5Inventory counting and basic stock tracking could be partially automated with barcode/RFID scanning systems and inventory management software, but the physical handling, verification of supply condition, and dynamic restocking decisions in a clinical setting require significant human oversight and discretion.
Task automatabilityclaude-sonnet-53/5Inventory tracking of clinical supplies is a data-entry/counting task that could largely be automated via RFID, barcode scanning, and inventory management software, though physical counting/restocking still requires human action.rating reflects partial automation potential.
Adoption barriersclaude-haiku-4-5-202510012/5Healthcare facilities have compliance and accountability requirements around medical supply tracking, but these do not legally mandate human performance of the inventory task itself—only verification and responsibility, which allows for automation with oversight.
Adoption barriersclaude-sonnet-52/5No licensing or legal requirement mandates a technologist perform inventory counts; it's largely an administrative task with low regulatory barriers, though hospital procurement systems may create organizational friction.
Cost vs. human wageclaude-haiku-4-5-202510012/5Current AI-powered inventory systems (software + integration + oversight) still cost substantially more than the direct labor of a technician performing manual spot checks and restocking in a small to mid-sized radiology department.
Cost vs. human wageclaude-sonnet-53/5Inventory software licensing and hardware (scanners, RFID tags) have upfront and maintenance costs comparable to the marginal labor time saved, especially for smaller imaging departments where volumes don't justify system costs.
Technical feasibility todayclaude-haiku-4-5-202510013/5Inventory management software and automated tracking systems exist in production healthcare settings, but they typically require human verification, manual correction of discrepancies, and judgment about supply adequacy—not fully autonomous end-to-end performance.
Technical feasibility todayclaude-sonnet-53/5Automated inventory management systems (barcode/RFID-based) are deployed in many hospitals for supply tracking, but MRI-specific consumable tracking often still relies on manual checks and spreadsheets in smaller facilities.

Write reports or notes to summarize testing procedures or outcomes for physicians or other medical professionals.

36

CI 25–48 · exposure 38 · augmentation 75 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Radiology and medical imaging remain moderately digitized but conservative in adoption of autonomous documentation AI. Most departments use transcription assistance or structured templates with human authorship; true AI-driven autonomous note generation is still in pilot phases at leading centers, not mainstream production.
Sector adoption velocityclaude-sonnet-52/5Healthcare documentation AI is spreading but adoption in imaging technologist workflows specifically remains slow due to EHR integration complexity and compliance caution.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-assisted documentation—such as AI-generated drafts, auto-populated structured fields, or real-time suggestions—significantly accelerates technologist note-writing and reduces cognitive load. Technologists retain full control and verification, making this a high-value augmentation scenario where AI can double productivity while humans remain accountable.
Augmentation potentialclaude-sonnet-54/5AI scribing and templated note generation can meaningfully speed up drafting of testing summaries, letting technologists focus on verification and edge cases.
Task automatabilityclaude-haiku-4-5-202510012/5AI can generate structured text summaries from MRI data and procedures, but the task requires translating complex radiological findings into clinically accurate narratives that match physician expectations and legal standards. Current systems cannot reliably extract, contextualize, and synthesize the nuanced clinical significance required without expert review, making end-to-end automation with 50% time savings unachievable.
Task automatabilityclaude-sonnet-53/5AI can draft structured summary notes from scan protocols and technologist observations, but capturing patient-specific technical nuances and quality issues still requires human verification, limiting full end-to-end automation.
Adoption barriersclaude-haiku-4-5-202510014/5Medical documentation is legally and clinically regulated; notes become part of the clinical record and may be used in litigation or licensing disputes. Physician oversight and sign-off are typically required by standard of care and institutional policy, creating a hard requirement that a human (often the technologist or radiologist) must validate and be accountable for the final note.
Adoption barriersclaude-sonnet-53/5Documentation must be accurate for patient safety and billing, and institutional/regulatory recordkeeping standards create moderate friction, though no strict licensing barrier prevents AI-assisted drafting.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI inference is cheap, the integration cost (EHR connectivity, compliance frameworks), quality assurance oversight by technologists, and physician review time add substantially to the actual per-task cost. These costs often approach or exceed the marginal cost of a technologist writing the note, especially in lower-volume settings.
Cost vs. human wageclaude-sonnet-53/5AI drafting could cut time spent on note writing, but integration, EHR compliance, and required human review keep costs roughly comparable rather than dramatically cheaper.
Technical feasibility todayclaude-haiku-4-5-202510013/5Clinical documentation tools and AI-powered transcription/summarization products exist and are used in some radiology departments, but they typically require significant technologist review and refinement. Error rates around clinical accuracy, missing details, and formatting inconsistencies remain material, limiting deployment to limited-scope or supervised-only scenarios rather than full independent operation.
Technical feasibility todayclaude-sonnet-52/5Some radiology reporting and clinical documentation assistants exist, but MRI technologist-specific procedural notes are not a widely deployed, mature product category distinct from radiologist reporting tools.

Take brief medical histories from patients.

29

CI 25–34 · exposure 25 · augmentation 63 · importance 4.7/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of autonomous patient intake AI remains pilot-stage; most MRI facilities use template-based forms or live technologist interviews, not deployed AI agents for history collection.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially imaging departments, is a historically slow-adopting sector for patient-facing automation due to safety and compliance concerns, though digital intake forms are becoming common.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could assist by pre-filling known medical information, highlighting missing critical fields, or transcribing patient speech, thereby reducing technologist data-entry burden and speeding intake while the human remains responsible for clinical accuracy.
Augmentation potentialclaude-sonnet-54/5AI-assisted intake forms and pre-visit questionnaires can efficiently gather and flag relevant history for the technologist to quickly review and verify, improving throughput.
Task automatabilityclaude-haiku-4-5-202510012/5Taking medical histories requires clarifying follow-up questions based on patient responses, distinguishing relevant from irrelevant information, and handling unexpected health conditions. While AI could fill in templated history forms with pre-written patient narratives, real-time adaptive questioning and clinical judgment needed for accuracy fall short of the 50% time-saving threshold today.
Task automatabilityclaude-sonnet-52/5Structured intake questions could be captured via chatbot or form, but clinical judgment about follow-up questions, patient anxiety, and contraindication screening (e.g., implants, pregnancy) still typically requires trained staff interaction.ureau.io
Adoption barriersclaude-haiku-4-5-202510014/5Medical history-taking is regulated under clinical standards and liability frameworks; documentation must be legally defensible and traceable to informed consent. Most healthcare systems retain human responsibility for accuracy, creating strong organizational and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement mandates a human ask these questions, but liability around missed contraindications (e.g., metal implants, pacemakers) creates strong practical incentive for human verification.
Cost vs. human wageclaude-haiku-4-5-202510012/5A technologist earns $25–35/hour loaded; current AI intake systems (including integration, data validation, and human oversight) cost comparably or more per patient encounter, offering no clear cost advantage.
Cost vs. human wageclaude-sonnet-53/5Digital intake forms are cheap to run, but the residual need for a trained technologist to verify and follow up on safety-critical answers keeps overall cost savings moderate rather than dramatic.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed clinical product reliably obtains complete, legally defensible medical histories autonomously from patients. Voice-based intake systems exist but require human review and correction, falling short of production-grade reliability for this safety-sensitive task.
Technical feasibility todayclaude-sonnet-52/5Some clinics use digital intake forms or kiosks for history-taking, but few production systems fully replace the technologist's verbal screening for MRI safety contraindications.

Review physicians' orders to confirm prescribed exams.

29

CI 25–34 · exposure 33 · augmentation 63 · importance 5.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare AI adoption remains cautious and heavily regulated; radiology facilities are typically risk-averse about automating clinical decision gates, and few have deployed end-to-end order-confirmation systems. Adoption is in the pilot or evaluation phase rather than production scale.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging is a highly regulated, moderately digitized sector where AI adoption for clinical workflow tasks like order verification remains largely pilot-stage rather than widespread production use.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully flag incomplete orders, suggest relevant protocols, or highlight inconsistencies, allowing technologists to review more efficiently; however, the core confirmation task still relies on human judgment and accountability, making augmentation meaningful but not transformative.
Augmentation potentialclaude-sonnet-54/5AI-based decision support can flag missing information, protocol mismatches, or contraindications, meaningfully speeding up the technologist's review while the human retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5While AI can parse and flag missing information in physician orders, confirming prescribed exams requires interpretation of clinical context, validation against patient history, and judgment about protocol appropriateness—tasks that still require human expertise and accountability. Current systems could assist with flagging obvious errors but cannot reliably perform the full confirmation task end-to-end.
Task automatabilityclaude-sonnet-53/5Confirming an order against protocol/patient data is a structured verification task AI can partially handle (matching order codes, flagging inconsistencies), but full confirmation often requires clinical judgment about contraindications and appropriateness that current systems only partially cover.'
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory oversight (FDA, state licensure), liability asymmetry (errors in confirming orders can harm patients), and the implicit requirement that a licensed technologist or radiologist verify orders before scanning create substantial legal and professional barriers to full automation.
Adoption barriersclaude-sonnet-54/5Medical order verification carries liability and regulatory implications; licensed technologists or radiologists must ultimately confirm exam appropriateness, creating a strong human sign-off requirement.
Cost vs. human wageclaude-haiku-4-5-202510012/5The cost of building and maintaining specialized clinical validation AI, including oversight infrastructure, likely approaches or exceeds the cost of a technologist performing this review task, given the low per-order automation savings and high liability risk if errors occur.
Cost vs. human wageclaude-sonnet-52/5Deploying validated clinical decision-support software requires integration, compliance, and ongoing human oversight, keeping costs closer to comparable rather than order-of-magnitude cheaper than technologist time for this narrow task.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some radiology IT systems can perform basic order validation (checking for completeness, patient demographics), but no deployed product reliably confirms the clinical appropriateness of prescribed exams or catches nuanced protocol mismatches without human review. Existing solutions require substantial human oversight.
Technical feasibility todayclaude-sonnet-52/5Some clinical decision-support and order-checking systems exist in radiology (e.g., appropriateness criteria checkers), but they are narrow-scope tools requiring human sign-off, not fully autonomous order confirmation in production.

Inspect images for quality, using magnetic resonance scanner equipment and laser camera.

28

CI 25–30 · exposure 30 · augmentation 63 · importance 4.6/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare AI adoption, while growing, remains relatively cautious in clinical-critical roles; MRI departments are typically conservative adopters bound by accreditation standards, and widespread production deployment of autonomous quality inspection in MRI centers is sparse compared to information-sector automation.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging is a highly regulated, slower-adopting sector; AI is used for image analysis/diagnosis pilots but real-time scan quality inspection by AI remains uncommon in production workflows.
Augmentation potentialclaude-haiku-4-5-202510014/5AI-powered image analysis tools that flag potential quality issues for technologist review can significantly improve speed and consistency of inspection workflows, allowing technologists to focus on interpretation and decision-making rather than manual scanning of every frame.
Augmentation potentialclaude-sonnet-53/5AI-based artifact detection and automated protocol adjustment tools can help technologists catch issues faster, offering moderate assistance while the technologist retains final judgment.
Task automatabilityclaude-haiku-4-5-202510012/5Quality inspection of MRI images requires detecting subtle artifacts, motion errors, and signal degradation—areas where AI shows promise but must still operate under human oversight due to high false-negative costs in clinical settings. While AI models can flag obvious defects, the task's demand for reliable, end-to-end automation at clinical-grade accuracy without human review remains unmet by deployed systems.
Task automatabilityclaude-sonnet-52/5Basic image quality checks (artifact detection, positioning, exposure) can be partially aided by software, but the full task of operating equipment, adjusting acquisition, and confirming diagnostic adequacy still requires a human technologist making judgment calls in real time.'
Adoption barriersclaude-haiku-4-5-202510014/5Medical imaging is heavily regulated (FDA oversight of diagnostic equipment and AI clinical decision support); quality control is often part of accreditation requirements (AABB, ACR standards); and clinical liability for missed defects creates strong disincentive to remove human technologists from the loop. Regulatory and liability barriers substantially protect this task.
Adoption barriersclaude-sonnet-54/5MRI operation and quality assurance is tightly regulated, requires licensed/certified technologists, and errors carry direct patient safety and diagnostic liability consequences.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference for image analysis is cheap, but integration into MRI workflows, validation against radiology standards, and the required human oversight add substantial overhead, keeping total cost closer to technologist wages than a clear cost advantage.
Cost vs. human wageclaude-sonnet-52/5AI QA tools add software licensing cost on top of the still-necessary technologist, so total cost is not clearly lower than having a trained tech perform the check directly.
Technical feasibility todayclaude-haiku-4-5-202510013/5AI-assisted image quality tools exist and are used in research and some clinical workflows, but they typically function as screening or flagging systems requiring technologist confirmation rather than fully autonomous inspection. Production deployments remain narrow and require material human oversight, falling short of mature, reliability-at-scale standards.
Technical feasibility todayclaude-sonnet-52/5Some vendor QA software flags obvious artifacts, but no deployed product autonomously inspects MRI images for clinical quality and decides whether to rescan without technologist review.

Attach physiological monitoring leads to patient's finger, chest, waist, or other body parts.

27

CI 5–49 · exposure 36 · augmentation 25 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare, especially radiology/imaging, has lagged in adopting robotic clinical task automation due to regulatory caution, liability aversion, and preference for human technologist presence and responsiveness. No significant production adoption of lead-attachment robots is evident in the sector.
Sector adoption velocityclaude-sonnet-51/5Healthcare imaging support roles involving direct physical patient contact show minimal AI-driven automation adoption for this specific physical task.
Augmentation potentialclaude-haiku-4-5-202510012/5AI augmentation is limited here since lead attachment is mechanical and procedural. AR/AI guidance for anatomical landmark identification might assist slightly, but the core task—physical placement and securing—offers minimal opportunity for AI-assisted productivity gain.
Augmentation potentialclaude-sonnet-52/5AI could potentially assist with monitoring data interpretation once leads are attached, but offers negligible assistance for the physical attachment process itself.
Task automatabilityclaude-haiku-4-5-202510015/5This task involves straightforward physical attachment of leads to standard body locations with minimal judgment or variation. Current robotic systems can reliably identify anatomical landmarks, position leads, and secure them, meeting the 50% time-saving threshold; human technologists spend significant time on lead placement and adjustment that could be fully automated.
Task automatabilityclaude-sonnet-51/5This is a physical, manual task requiring hands-on placement of leads on a patient's body, which current AI systems cannot perform without robotic embodiment, which is not deployed for this purpose.
Adoption barriersclaude-haiku-4-5-202510014/5Regulatory oversight (FDA, medical device classification), liability concerns over patient contact and skin contact safety, and hospital certification requirements create substantial friction. Patient comfort, rare skin sensitivities, and perceived safety risks also impose organizational and liability barriers to autonomous systems.
Adoption barriersclaude-sonnet-54/5Direct patient contact for medical monitoring setup typically requires trained, often licensed personnel, and errors in lead placement can affect diagnostic quality or patient safety, creating strong barriers to non-human performance.
Cost vs. human wageclaude-haiku-4-5-202510012/5A robotic system capable of safe, reliable lead placement would require significant capital investment ($500k+), integration, and maintenance costs that exceed the annual loaded wage of a single technologist. Per-task cost would be comparable or higher than human performance.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI-based substitute performing this physical task, so AI cost is not comparable; the human remains the only option, making AI effectively more expensive/infeasible.
Technical feasibility todayclaude-haiku-4-5-202510012/5While robotic arms exist in research and some healthcare labs, no mature deployed product systematically automates lead attachment at clinical scale in production MRI environments. Prototypes exist but lack the reliability, adaptability to patient variation, and regulatory clearance for routine clinical use.
Technical feasibility todayclaude-sonnet-51/5No deployed products perform physical lead placement on patients; this remains a manual clinical task performed by technologists.

Conduct screening interviews of patients to identify contraindications, such as ferrous objects, pregnancy, prosthetic heart valves, cardiac pacemakers, or tattoos.

26

CI 23–29 · exposure 25 · augmentation 50 · importance 5.0/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of autonomous AI for clinical screening tasks remains very slow due to regulatory caution, liability risk, and institutional preference for human clinical judgment on safety-critical decisions. Pilot projects exist, but production deployment in real MRI departments is rare.
Sector adoption velocityclaude-sonnet-52/5Healthcare, especially imaging safety protocols, is a slow-adopting, highly regulated sector where automation of safety-critical patient interviews has seen minimal real-world deployment.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist by presenting a structured questionnaire, flagging keywords in patient records that hint at contraindications, or organizing patient responses for the technologist to review. This augmentation reduces data entry burden but does not transform the core clinical judgment required.
Augmentation potentialclaude-sonnet-53/5AI-assisted digital forms and checklists can pre-screen patients and flag risks for the technologist to verify, meaningfully speeding up the interview process while keeping a human accountable.
Task automatabilityclaude-haiku-4-5-202510012/5While AI could assist in analyzing patient history documents or flagging keywords related to contraindications, the task requires real-time clinical judgment, follow-up questioning to clarify vague responses, and assessment of patient reliability that current AI cannot reliably perform end-to-end. The nuanced screening of complex medical histories and patient disclosure remains heavily dependent on human clinical expertise.
Task automatabilityclaude-sonnet-52/5While intake questionnaires could be digitized and AI could flag risk keywords, the task requires physical verification, follow-up probing, and judgment about ambiguous or incomplete patient responses that current systems cannot fully replace end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5Strong barriers exist: patient safety is paramount, liability for missed contraindications falls on the facility and supervising physician, and regulatory bodies expect qualified clinical personnel to conduct and sign off on screening. Many imaging centers legally require a technologist or radiologist to review and validate the screening before proceeding.
Adoption barriersclaude-sonnet-54/5Missed contraindications (e.g., pacemakers, ferrous implants) can cause serious injury or death, creating strong liability and regulatory pressure for a trained human to perform or verify this screening.
Cost vs. human wageclaude-haiku-4-5-202510012/5An AI-assisted questionnaire system might reduce technologist time modestly, but the cost of development, validation, integration with EHRs, and continuous oversight would likely exceed the savings from automating a 10–15 minute interview task, especially given liability concerns.
Cost vs. human wageclaude-sonnet-53/5A digital screening form is cheap to run, but liability concerns require human verification and sign-off, so the effective all-in cost including oversight is comparable to having a technologist do it directly.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product reliably performs comprehensive MRI screening interviews independently. Chatbots and questionnaire systems exist but lack the ability to conduct genuine clinical assessment, probe ambiguous answers, or make judgment calls on borderline contraindications that would be required in production MRI environments.
Technical feasibility todayclaude-sonnet-52/5Some clinics use digital intake forms with basic logic branching, but no widely deployed AI product independently conducts full MRI safety screening interviews including verbal clarification and physical checks.

Select appropriate imaging techniques or coils to produce required images.

25

CI 25–25 · exposure 25 · augmentation 50 · importance 4.9/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5MRI facilities operate in conservative, highly regulated healthcare environments with slow technology adoption cycles. Most imaging departments still rely on manual protocol selection and radiologist feedback rather than AI-assisted or autonomous systems.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging is a highly regulated, moderately digitized sector where AI protocol tools are emerging but full-scale autonomous adoption remains slow and cautious.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can usefully suggest protocols based on procedure type and patient history, helping technologists work faster and reducing cognitive load on complex cases. However, the human technologist must validate and adjust recommendations, making this genuine assistance rather than full automation.
Augmentation potentialclaude-sonnet-53/5AI-based protocol recommendation and auto-positioning tools can meaningfully speed up technologist decision-making while the technologist remains responsible for final selection.
Task automatabilityclaude-haiku-4-5-202510012/5Selecting imaging techniques and coils requires understanding patient anatomy, pathology, and clinical questions—expert judgment that current AI cannot reliably perform end-to-end. While AI can assist in protocol suggestion, the final selection demands real-time assessment of patient factors and scanner constraints that exceed today's autonomous capability.
Task automatabilityclaude-sonnet-52/5Selecting imaging protocols and coils requires clinical judgment based on patient anatomy, contraindications, and diagnostic goals that current AI cannot reliably synthesize end-to-end without human oversight.assistant
Adoption barriersclaude-haiku-4-5-202510014/5MRI protocol selection is tightly regulated by FDA, ACR, and institutional protocols; liability for suboptimal image quality or patient safety rests on qualified personnel. Regulatory and liability frameworks strongly protect human technologist decision-making authority.
Adoption barriersclaude-sonnet-54/5Patient safety, coil/contraindication management, and regulatory requirements mean a licensed MRI technologist must make and be accountable for these decisions.
Cost vs. human wageclaude-haiku-4-5-202510012/5Integration of AI decision-support into MRI workflows requires custom development, validation, and ongoing clinical oversight—costs that approach or exceed the time savings from faster protocol selection by a technologist earning $35–50k annually.
Cost vs. human wageclaude-sonnet-52/5AI-assisted protocol suggestion software adds licensing and integration cost while still requiring a paid technologist to verify and execute the choice, so cost savings are limited.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed product independently selects MRI protocols and coils in clinical production. Research prototypes and decision-support tools exist but require significant technologist oversight and validation; they operate as narrow assistants rather than autonomous systems.
Technical feasibility todayclaude-sonnet-52/5Some MRI systems offer protocol suggestion or auto-selection features, but these are decision-support tools requiring technologist verification, not autonomous production systems.

Explain magnetic resonance imaging (MRI) procedures to patients, patient representatives, or family members.

25

CI 16–34 · exposure 17 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare adoption of AI for direct patient-facing clinical communication remains limited; hospitals prioritize human technologists for patient interaction due to liability and quality-of-care concerns, and compliance frameworks have not yet normalized AI as a substitute for technologist-delivered patient education.
Sector adoption velocityclaude-sonnet-52/5Healthcare adoption of AI for direct patient communication is slower than in tech/finance sectors due to compliance, trust, and workflow integration challenges, though patient education chatbots are slowly emerging.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist technologists by generating procedure explanations, pre-populating educational materials, or suggesting answers to common questions, reducing routine explanation time, but the technologist must verify and deliver explanations to maintain patient trust and ensure adequate comprehension.
Augmentation potentialclaude-sonnet-54/5AI can generate consistent, multilingual, and easy-to-understand explanatory scripts, videos, or FAQs that technologists can use or supplement during patient interactions, meaningfully improving efficiency and consistency.
Task automatabilityclaude-haiku-4-5-202510011/5Explaining MRI procedures requires dynamic interaction with anxious patients, assessing comprehension in real-time, and adapting explanations to individual concerns and literacy levels—tasks that demand genuine empathy and contextual responsiveness that current AI systems cannot reliably execute end-to-end in clinical settings.
Task automatabilityclaude-sonnet-52/5AI can generate patient-facing explanations of MRI procedures (chatbots, printed materials), but the task also involves live, adaptive communication addressing patient anxiety, contraindications, and specific clinical context that requires human presence and judgment.
Adoption barriersclaude-haiku-4-5-202510014/5Significant barriers exist: patient safety liability if explanations are inadequate or miscommunicated, professional standards expecting human technologists to directly manage patient anxiety and consent, and regulatory expectations that qualified staff verify patient understanding before MRI procedures.
Adoption barriersclaude-sonnet-53/5No strict licensing requirement mandates a human explain the procedure, but patient safety screening (metal implants, contrast allergies, claustrophobia) is bundled into this interaction, creating clinical and liability reasons for human involvement.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI inference cost for explanation generation is low, but integration into clinical workflows, human oversight for safety/liability, and the need for fallback to human technologists for complex patient interactions make total cost-per-adequate-outcome comparable to or higher than human staff time.
Cost vs. human wageclaude-sonnet-53/5Static AI-generated educational content is cheap to produce and reuse, but the in-person explanation component still requires the technologist's time, so overall cost savings are only partial.
Technical feasibility todayclaude-haiku-4-5-202510012/5While AI chatbots can generate generic MRI procedure explanations, no deployed product reliably handles the nuanced, real-time interpersonal communication this task demands, including reassurance of anxious patients, assessment of understanding, and handling of sensitive concerns in a clinical context.
Technical feasibility todayclaude-sonnet-52/5Deployed patient-education chatbots and pre-visit informational videos exist in some hospital systems, but real-time, personalized explanation immediately before the procedure is still performed by the technologist in virtually all facilities.

Instruct medical staff or students in magnetic resonance imaging (MRI) procedures or equipment operation.

21

CI 16–25 · exposure 17 · augmentation 63 · importance 4.4/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Hospitals and medical institutions adopt AI for diagnostic support and operational efficiency, but training and instruction remain human-centric domains with slow adoption of automated teaching. Supplementary AI tools (e.g., eLearning) are used, but human instructors remain the standard.
Sector adoption velocityclaude-sonnet-52/5Healthcare training programs are adopting AI-assisted e-learning tools slowly, with clinical/physical instruction remaining a laggard area due to hands-on and safety requirements.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can assist instructors by generating preliminary training materials, drafting procedure guides, creating practice quizzes, or providing on-demand reference content, but the core instructional relationship and real-time feedback remain human-dependent. Useful support exists but does not transform the teaching process fundamentally.
Augmentation potentialclaude-sonnet-54/5AI can meaningfully augment this task by generating training materials, quizzes, simulations, and explanatory content that instructors use to supplement hands-on teaching.
Task automatabilityclaude-haiku-4-5-202510011/5Instructing medical staff or students requires real-time interaction, adaptive feedback based on learner comprehension, demonstration of hands-on equipment operation, and personalized guidance—capabilities that current AI systems cannot reliably deliver at parity with human instruction. The task involves responsive teaching, which falls outside what deployed AI can perform end-to-end.
Task automatabilityclaude-sonnet-52/5Teaching hands-on MRI procedures and equipment operation involves physical demonstration, supervised practice, and real-time correction that current AI cannot perform end-to-end, though some lecture content could be automated.
Adoption barriersclaude-haiku-4-5-202510014/5Medical education and training of clinical staff are subject to accreditation standards, regulatory oversight (e.g., radiology board requirements), and liability concerns; institutions have strong preferences for credentialed instructors to sign off on competency. Legal and organizational requirements create substantial friction against AI-only instruction.
Adoption barriersclaude-sonnet-54/5Clinical training often requires certified technologists or radiologic educators to sign off on competency, and licensing/accreditation bodies typically mandate qualified human instruction for hands-on MRI operation.
Cost vs. human wageclaude-haiku-4-5-202510012/5While AI-generated instructional content (videos, modules) can reduce some material costs, live instruction by qualified technologists involves significant human expertise and liability; AI tools would require substantial human oversight and customization, making the all-in cost comparable to or higher than direct human instruction for rigorous medical training.
Cost vs. human wageclaude-sonnet-52/5While AI-generated training materials are cheap, the hands-on supervision and certification components still require paid expert instructors, keeping overall cost comparable to or higher than partial AI substitution.
Technical feasibility todayclaude-haiku-4-5-202510012/5Some narrow products (e.g., video tutorials, chatbot FAQs) can deliver pre-scripted MRI information, but no deployed system can serve as a reliable primary instructor for complex procedural knowledge, safety-critical protocols, or hands-on equipment training at the level expected in medical education.
Technical feasibility todayclaude-sonnet-52/5AI tutoring products exist for medical education content delivery, but no deployed product reliably instructs hands-on MRI equipment operation or supervises clinical training in production settings.

Troubleshoot technical issues related to magnetic resonance imaging (MRI) scanner or peripheral equipment, such as monitors or coils.

21

CI 16–25 · exposure 20 · augmentation 50 · importance 4.2/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510012/5Healthcare IT adoption is moderate, and MRI troubleshooting remains tightly coupled to on-site human expertise. While some hospitals use remote monitoring and diagnostic software, autonomous or AI-driven troubleshooting has not achieved significant production deployment in the sector.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging equipment maintenance is a physically-oriented, highly regulated niche with slow AI integration compared to purely digital workflows.
Augmentation potentialclaude-haiku-4-5-202510013/5AI can meaningfully assist by analyzing error logs, suggesting likely causes, and retrieving relevant service documentation, thereby reducing diagnostic time for technicians. However, the task's heavy reliance on physical inspection and hands-on repair limits transformative augmentation to the diagnostic phase.
Augmentation potentialclaude-sonnet-53/5AI-driven diagnostic software and knowledge bases can help technologists quickly identify likely fault codes or guide troubleshooting steps, improving efficiency without replacing hands-on work.
Task automatabilityclaude-haiku-4-5-202510012/5MRI troubleshooting requires physical diagnostics, hands-on equipment inspection, and real-time problem-solving in complex electromechanical systems. While AI could assist in symptom classification and diagnostic suggestions, the task inherently demands on-site inspection, specialized hardware knowledge, and real-time intervention that current AI cannot perform end-to-end without substantial human oversight.
Task automatabilityclaude-sonnet-52/5Requires physical inspection, hands-on diagnostics, and manipulation of hardware (coils, cables, monitors) that current AI cannot perform end-to-end; AI can assist with diagnostic checklists or error code interpretation but not full remediation.'
Adoption barriersclaude-haiku-4-5-202510014/5MRI systems are medical devices subject to FDA regulation and manufacturer warranty constraints. Troubleshooting and repairs often require certified technicians and must comply with equipment-vendor protocols. Liability for equipment failure and patient safety create strong regulatory and contractual barriers to full automation.
Adoption barriersclaude-sonnet-54/5MRI equipment is safety-critical and regulated; service and troubleshooting often require certified personnel or vendor-authorized technicians, creating strong barriers to full automation.
Cost vs. human wageclaude-haiku-4-5-202510012/5MRI technicians earn substantial salaries ($60k–$75k+ loaded), and troubleshooting is a high-skill task. AI tools for diagnostic support are relatively cheap, but cannot replace the on-site technician; the total cost of AI + human oversight would rival or exceed the human-alone cost for most scenarios.
Cost vs. human wageclaude-sonnet-51/5AI cannot substitute the physical labor and specialized service calls involved, so no cost savings are realized versus trained technologists or field engineers.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI product reliably troubleshoots MRI hardware faults autonomously. AI can support diagnosis through data analysis and documentation review, but actual troubleshooting—identifying failed components, recalibrating systems, replacing parts—requires human technicians. Existing systems are in research or early pilot phases.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously diagnoses and resolves MRI hardware faults today; troubleshooting remains a manual technician/service-engineer task, sometimes aided by manufacturer software logs.

Test magnetic resonance imaging (MRI) equipment to ensure proper functioning and performance in accordance with specifications.

15

CI 5–25 · exposure 13 · augmentation 50 · importance 4.5/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare settings, particularly medical imaging departments, have been slow to automate core technical functions due to regulatory constraints, safety criticality, and preference for certified human oversight. Adoption of AI in equipment testing remains minimal and experimental rather than widespread production deployment.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging operations adopt software-based diagnostics and monitoring tools gradually, but hands-on equipment testing tasks show little evidence of AI-driven displacement in practice.
Augmentation potentialclaude-haiku-4-5-202510013/5AI could usefully assist technologists by automating data logging, flagging anomalies in test results, or predicting maintenance needs from historical performance patterns, thereby raising productivity while the technologist retains judgment on corrective actions and compliance decisions.
Augmentation potentialclaude-sonnet-53/5AI-enabled diagnostic software embedded in MRI systems can flag anomalies and support technologists in identifying performance issues, but the physical testing and verification remain human-led.
Task automatabilityclaude-haiku-4-5-202510012/5Testing MRI equipment requires hands-on calibration, physical setup adjustments, and real-time sensor monitoring that current AI cannot perform autonomously. While AI could assist with data analysis of test outputs, the core task of ensuring hardware compliance through manual intervention and troubleshooting remains primarily human-dependent.
Task automatabilityclaude-sonnet-51/5This requires physical inspection, hands-on calibration checks, and manipulation of MRI hardware, which current AI systems cannot perform end-to-end without robotic embodiment.
Adoption barriersclaude-haiku-4-5-202510014/5Medical device testing is heavily regulated; MRI technologists are licensed professionals, and equipment performance verification often requires signed certification for regulatory compliance. Healthcare liability and the requirement for qualified personnel sign-off create substantial legal and organizational barriers to full automation.
Adoption barriersclaude-sonnet-54/5Medical equipment quality assurance is subject to safety regulations, manufacturer specifications, and institutional accreditation requirements that typically mandate qualified personnel to perform and document these checks.
Cost vs. human wageclaude-haiku-4-5-202510012/5AI-based solutions for MRI equipment testing would require significant integration with specialized hardware and systems, likely making per-test costs comparable to or higher than a trained technologist's labor, especially accounting for liability and oversight requirements.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing this physical equipment-testing task, so no meaningful cost comparison favors AI; a human technologist is required regardless of cost.
Technical feasibility todayclaude-haiku-4-5-202510012/5No deployed AI product reliably performs full MRI equipment testing independently. AI systems exist for analyzing MRI images and some diagnostic support, but equipment functional testing, calibration verification, and compliance checking require human technologists and specialized domain knowledge in production settings.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product independently tests or physically verifies MRI equipment functioning; this remains a human technologist task supported at most by built-in diagnostic software.

Provide headphones or earplugs to patients to improve comfort and reduce unpleasant noise.

9

CI 0–18 · exposure 0 · augmentation 0 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5This routine clinical support task is performed at the point of care by technicians; there is no incentive or path for AI adoption in healthcare settings due to regulatory and physical constraints.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging settings adopt AI mainly for image analysis and workflow, not for physical patient-comfort tasks like distributing earplugs.
Augmentation potentialclaude-haiku-4-5-202510011/5AI offers no meaningful assistance in selecting or fitting audio equipment for patient comfort during an MRI scan.
Augmentation potentialclaude-sonnet-51/5AI provides no meaningful assistance for this simple physical task of distributing hearing protection to patients.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical interaction with patients (fitting headphones or earplugs) and judgment about patient comfort, necessitating human presence and dexterity that current AI systems cannot perform end-to-end.
Task automatabilityclaude-sonnet-51/5This is a simple physical act of handing protective equipment to a patient, requiring physical presence and manual dexterity that current AI systems cannot perform.'
Adoption barriersclaude-haiku-4-5-202510015/5Direct patient contact and care activities are legally and operationally restricted to licensed healthcare staff; regulatory and liability frameworks require human personnel to manage patient comfort during medical procedures.
Adoption barriersclaude-sonnet-52/5No licensing requirement specifically governs this minor action, though it occurs within a regulated clinical environment requiring human presence for patient care and safety.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of deploying a robotic system to place headphones on patients would far exceed the minimal labor cost of a technician performing this simple task.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute for this trivial manual action, so any AI-based approach (e.g. robotics) would be far more expensive than a human simply handing over earplugs.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product can autonomously provide personal protective equipment or comfort devices to patients; this requires a physical technician or staff member present.
Technical feasibility todayclaude-sonnet-51/5No deployed AI product or robot performs this simple physical patient-comfort task in clinical MRI settings today.

Operate magnetic resonance imaging (MRI) scanners.

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, particularly imaging, adopts automation slowly. MRI operation remains predominantly manual in all major healthcare systems; no evidence of significant displacement or agent-based automation in production settings.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging is a moderately digitized but highly regulated, hands-on physical sector where automation of scanner operation itself is not underway, though AI adoption in image analysis is progressing.
Augmentation potentialclaude-haiku-4-5-202510013/5AI assists with image reconstruction, quality assessment, and protocol recommendations, helping technologists work more efficiently. However, the core task of operating the scanner and managing patient safety remains human-led, limiting augmentation to supporting functions.
Augmentation potentialclaude-sonnet-53/5AI can assist with scan protocol selection, image quality checks, and workflow optimization, but the hands-on scanner operation and patient positioning remain manual.
Task automatabilityclaude-haiku-4-5-202510012/5Operating MRI scanners requires real-time physical manipulation of equipment, patient positioning, monitoring of physiological responses, and complex troubleshooting of hardware failures. While AI can assist with image reconstruction and some protocol optimization, the hands-on operation and safety-critical decision-making during scans cannot be fully automated today.
Task automatabilityclaude-sonnet-51/5Physically operating an MRI scanner requires manipulating patients, positioning coils, handling contrast injections, and responding to real-time physical events—none of which current AI systems can perform without embodiment.
Adoption barriersclaude-haiku-4-5-202510015/5MRI operation is subject to strict FDA and state medical device regulations, requires state licensure/certification as a radiologic technologist, and involves direct patient safety responsibility. Liability for equipment malfunction or patient harm creates powerful legal and regulatory barriers to full automation.
Adoption barriersclaude-sonnet-55/5Operating MRI equipment requires licensed technologist certification, direct patient contact, safety screening for implants/metal, and regulatory oversight, making substitution legally and physically infeasible.
Cost vs. human wageclaude-haiku-4-5-202510011/5The capital and integration costs of automating MRI hardware control, plus the liability and oversight infrastructure required, far exceed the cost of employing a technologist. Automation is not economically viable given the complexity and safety requirements.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing the physical operation task, so cost comparison favors the human by default since no AI alternative exists to price.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system operates MRI scanners end-to-end in production. MRI operation demands continuous physical presence, real-time patient interaction, and hardware control that requires human operators; no commercial product substitutes for this.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously operates MRI hardware end-to-end; AI is used for image reconstruction/analysis, not scanner operation and patient handling.

Connect physiological leads to physiological acquisition control (PAC) units.

5

CI 5–5 · exposure 0 · augmentation 25 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a sector with slow automation adoption for patient-facing tasks. Hospital workflows and regulatory requirements make rapid substitution of equipment setup tasks unlikely.
Sector adoption velocityclaude-sonnet-51/5Healthcare physical/manual tasks show very slow AI adoption; imaging technologist roles involving direct patient contact remain highly resistant to automation.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially guide technologists through procedural steps or verify correct connections via vision systems, but the task itself is too manual and brief to yield significant productivity gains from augmentation.
Augmentation potentialclaude-sonnet-52/5AI could potentially assist with checklists, lead-placement guidance displays, or verifying correct connection via sensor feedback, but this offers only marginal support to the core physical task.
Task automatabilityclaude-haiku-4-5-202510011/5Connecting physiological leads to PAC units is a fine-motor, spatial task requiring precise physical manipulation in a clinical setting. Current AI systems lack the embodied dexterity and real-time sensorimotor feedback needed to reliably handle medical equipment connectors.
Task automatabilityclaude-sonnet-51/5This is a physical manipulation task requiring hands to attach ECG/respiratory leads to a patient's body, which current AI systems cannot perform without embodied robotics.-Deployed general-purpose robots for this exist nowhere in clinical practice today.
Adoption barriersclaude-haiku-4-5-202510014/5Clinical settings have strict protocols around equipment setup and patient safety. Connection errors could compromise data quality or patient monitoring, creating liability and regulatory oversight barriers. Institutional practices typically require trained personnel to verify setup.
Adoption barriersclaude-sonnet-54/5Requires direct physical patient contact and clinical judgment about lead placement in a medical imaging environment, with safety and liability concerns around patient monitoring accuracy.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of robotic systems capable of performing delicate medical equipment assembly, including hardware, maintenance, and integration, far exceeds the wage cost of a technologist to perform this brief task.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI/robotic substitute, so the human technologist remains the only cost-effective option for this physical task.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product today can autonomously connect physiological leads to equipment with the consistency required in clinical practice. This remains a task requiring human hands-on work.
Technical feasibility todayclaude-sonnet-51/5No commercial product performs physical patient lead attachment in MRI settings; this remains a manual clinical task performed by trained technologists.

Calibrate magnetic resonance imaging (MRI) console or peripheral hardware.

4

CI 0–7 · exposure 0 · augmentation 25 · importance 4.3/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare MRI operations remain heavily dependent on certified human technologists for all calibration work. No sector-wide AI automation of this task is evident; healthcare digitization in this domain is limited to data management, not autonomous hardware calibration.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging is a highly regulated, physically-grounded sector with slower AI deployment for hands-on equipment tasks compared to purely digital/information work.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could marginally assist by analyzing historical calibration data or flagging drift patterns, but the core physical and decision-making work remains firmly with the technologist. Any augmentation is limited to diagnostic support rather than meaningful productivity transformation.
Augmentation potentialclaude-sonnet-52/5Some diagnostic software and automated quality-control checks can flag calibration issues or guide technologists, but this offers only modest assistance to the core physical calibration task.
Task automatabilityclaude-haiku-4-5-202510011/5MRI calibration requires hands-on physical intervention with complex hardware, real-time adjustment based on equipment-specific responses, and expert troubleshooting of sensor readings. Current AI systems cannot autonomously perform the mechanical and electronic adjustments necessary to calibrate MRI consoles, nor can they supervise the real-time feedback loops required.
Task automatabilityclaude-sonnet-51/5Calibration requires physical interaction with hardware, coil placement, phantom testing, and troubleshooting equipment-specific quirks that current AI cannot perform end-to-end without a human physically present and operating controls.
Adoption barriersclaude-haiku-4-5-202510015/5MRI calibration is tightly regulated; technologists must be licensed and certified by regulatory bodies (FDA, state licensing boards). Liability for equipment malfunction and patient safety create hard legal barriers—a licensed human must legally perform or directly supervise this task.
Adoption barriersclaude-sonnet-54/5MRI equipment calibration involves patient safety, regulatory compliance (FDA-cleared devices), and typically requires certified technologists or service engineers, creating strong barriers to non-human substitution.
Cost vs. human wageclaude-haiku-4-5-202510011/5AI solutions for this task do not exist at any meaningful cost point. A certified MRI technologist performing calibration would far undercut any speculative automation cost, given the current absence of viable alternatives.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute performing this physical calibration task, so the cost comparison favors the human technologist who must be on-site regardless.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product today can independently calibrate MRI hardware. This task demands deep domain expertise, specialized equipment access, and immediate physical action—none of which existing AI systems handle in production environments.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously calibrates MRI console or peripheral hardware; this remains a manual technologist task performed with vendor service tools and manufacturer protocols, sometimes with software-assisted diagnostics but not full automation.

Operate optical systems to capture dynamic magnetic resonance imaging (MRI) images, such as functional brain imaging, real-time organ motion tracking, or musculoskeletal anatomy and trajectory visualization.

4

CI 0–7 · exposure 0 · augmentation 38 · importance 4.1/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare remains a laggard in AI adoption for clinical workflows; MRI departments have shown minimal adoption of autonomous scanning systems and continue to rely on certified technologists for acquisition. No evidence of displacement or production-scale automation in this domain.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging is a highly regulated, physically-grounded sector with slow AI adoption for hands-on operational tasks despite faster uptake in image analysis software.
Augmentation potentialclaude-haiku-4-5-202510012/5AI can assist with post-scan image reconstruction, artifact detection, and preliminary quality assessment, but does not meaningfully augment the live acquisition process itself. The technologist still performs the full operational task with limited AI assistance during scanning.
Augmentation potentialclaude-sonnet-53/5AI can assist with image reconstruction, motion correction, and protocol optimization during dynamic MRI acquisition, improving image quality and workflow efficiency for the technologist.
Task automatabilityclaude-haiku-4-5-202510011/5Operating MRI optical systems requires real-time hardware manipulation, troubleshooting scanner malfunction, positioning patient-specific equipment, and expert judgment about image quality—skills that current AI cannot perform end-to-end. The task involves continuous sensor feedback loops and physical interaction with complex equipment that exceed today's automation capabilities.
Task automatabilityclaude-sonnet-51/5This requires physically operating MRI equipment, positioning patients, and managing hardware/software in real time, which is a hands-on clinical task no current AI can perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510014/5MRI operation is tightly regulated by FDA and joint accreditation bodies; technologists typically require certification (ARRT-MR or equivalent) and licensing varies by jurisdiction. Medical liability, patient safety during scanning, and regulatory requirements for human oversight create substantial legal and organizational barriers to full automation.
Adoption barriersclaude-sonnet-55/5Operating imaging equipment on patients requires licensed technologist certification, regulatory compliance, and direct human presence for patient safety and equipment handling.
Cost vs. human wageclaude-haiku-4-5-202510011/5MRI systems are capital-intensive and require specialized facilities; AI cannot yet replace the technician's role in hardware operation, meaning automation cost-benefit does not yet apply. A technologist's labor is still far cheaper than attempting to automate the full operational workflow.
Cost vs. human wageclaude-sonnet-51/5There is no AI substitute performing the physical scanning task, so cost comparison favors the human technologist entirely; AI cannot replace the labor involved.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably operates MRI hardware independently or captures dynamic imaging sequences without human oversight. While AI assists with post-acquisition image reconstruction and analysis, the acquisition itself remains a human-performed task with no production alternatives.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously operates MRI scanners for dynamic imaging capture; this remains a technologist-driven physical and technical process.

Comfort patients during exams, or request sedatives or other medication from physicians for patients with anxiety or claustrophobia.

1

CI 0–3 · 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 facilities maintain strict staffing and licensing requirements for MRI technologists; there is minimal incentive or ability to automate patient comfort and medication coordination roles given regulatory constraints and patient safety imperatives.
Sector adoption velocityclaude-sonnet-52/5Healthcare imaging settings adopt AI for image analysis and scheduling but direct patient-facing comfort and medication-request interactions see essentially no automation adoption.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could potentially provide pre-exam anxiety information or chatbot support before the procedure, but offers limited augmentation during the actual exam since the technologist's primary value is in-person presence, communication, and clinical coordination with physicians.
Augmentation potentialclaude-sonnet-52/5AI could provide minor support like anxiety-reduction audio/visual aids or automated alerts flagging patient distress, but the core comforting and clinical decision-making remains human-driven with limited AI assistance.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires direct human interaction, empathetic communication, and clinical judgment to assess patient distress and coordinate with physicians. AI cannot provide in-person comfort, emotional reassurance, or make real-time clinical decisions about sedation needs.
Task automatabilityclaude-sonnet-51/5This is an interpersonal comforting task combined with a clinical judgment call to request medication, requiring physical presence, empathy, and real-time human interaction that current AI cannot replicate or automate end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5Strong legal and regulatory barriers exist: only licensed physicians can prescribe or authorize sedatives, and patient safety during MRI requires qualified human personnel present. The human contact requirement and liability for adverse reactions create hard barriers to automation.
Adoption barriersclaude-sonnet-55/5Requesting sedatives requires licensed clinical judgment and physician authorization, and direct patient care/comfort during a medical procedure requires human presence and accountability under healthcare regulations.
Cost vs. human wageclaude-haiku-4-5-202510011/5The task requires a licensed medical professional present during the exam; any AI solution would require human oversight and presence anyway, making the all-in cost equal to or higher than the current human-only approach.
Cost vs. human wageclaude-sonnet-51/5There is no AI system performing this task, so no meaningful cost comparison exists; a human technologist must be present regardless.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI system can autonomously comfort patients in-person or make prescribing recommendations to physicians. While chatbots might provide pre-exam anxiety information, they cannot replace the human presence and clinical assessment required during an actual MRI procedure.
Technical feasibility todayclaude-sonnet-51/5No deployed product provides patient comfort during medical imaging or makes clinical medication requests; this remains squarely a human clinical function.

Place and secure small, portable magnetic resonance imaging (MRI) scanners on body part to be imaged, such as arm, leg, or head.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare automation of physical patient-contact tasks lags far behind information-based sectors due to regulatory hurdles, liability concerns, and organizational conservatism in clinical settings.
Sector adoption velocityclaude-sonnet-51/5Healthcare imaging is a highly regulated, physically-grounded sector with minimal automation of hands-on patient positioning tasks; adoption of physical automation here is essentially nonexistent.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could assist with pre-scan planning or protocol selection, but offers minimal direct assistance during the physical placement and securing task itself, which remains dominated by human tactile skill and judgment.
Augmentation potentialclaude-sonnet-52/5AI may assist with imaging protocol selection or scan parameter optimization, but offers negligible help with the physical act of placing and securing equipment on a patient.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical manipulation of delicate medical equipment on a patient's body—positioning, securing, and ensuring comfort and safety. Current AI systems cannot perform end-to-end physical manipulation, patient handling, or real-time adjustment based on tactile and visual feedback in clinical settings.
Task automatabilityclaude-sonnet-51/5This is a physical manipulation task requiring hands-on placement and securing of equipment against a patient's body, which current AI systems and robotics cannot perform reliably or safely.
Adoption barriersclaude-haiku-4-5-202510015/5This task carries high regulatory and liability barriers: medical device handling is governed by FDA and clinical protocols; patient safety requires human judgment and accountability; direct patient contact creates legal and consent requirements that effectively mandate human oversight and sign-off.
Adoption barriersclaude-sonnet-55/5Positioning imaging equipment on patients requires licensed technologist judgment for patient safety, comfort, and diagnostic accuracy, with clear regulatory and liability requirements for human performance.
Cost vs. human wageclaude-haiku-4-5-202510011/5Deploying a robotic system capable of safe, precise placement would require significant capital investment, maintenance, and specialized integration far exceeding the cost of a trained MRI technologist performing this task.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI/robotic substitute performing this physical task, so the human technologist remains the only cost-effective option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product reliably performs this task in production. While robotics research explores device placement, no clinical system demonstrates the dexterity, safety validation, and liability clearance needed for autonomous or semi-autonomous MRI scanner placement on patients.
Technical feasibility todayclaude-sonnet-51/5No deployed product autonomously positions and secures portable MRI scanners on patients; this remains a manual clinical task performed by trained technologists.

Position patients on cradle, attaching immobilization devices, if needed, to ensure appropriate placement for imaging.

0

CI 0–0 · exposure 0 · augmentation 25 · importance 4.8/5 · click for rater detail

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare adoption of physical automation in patient-facing roles remains minimal; the sector is conservative due to liability, regulation, and strong preference for human care continuity. No evidence of production-level robotic patient positioning systems in hospitals.
Sector adoption velocityclaude-sonnet-51/5Healthcare physical patient-handling tasks show minimal AI/robotic adoption; this is a highly manual, low-digitization component of an otherwise digitizing field.
Augmentation potentialclaude-haiku-4-5-202510012/5AI tools could assist by recommending optimal positioning angles or immobilization setup based on imaging protocols, but the hands-on physical work of moving and securing patients will remain human-performed. Augmentation opportunity is narrow.
Augmentation potentialclaude-sonnet-52/5Some smart positioning guidance systems or sensors may assist technologists with alignment, but the core physical act of positioning and immobilizing patients sees limited AI augmentation today.
Task automatabilityclaude-haiku-4-5-202510011/5This task requires physical manipulation, spatial reasoning about patient comfort and safety, and real-time adjustment based on patient feedback—capabilities that current robotic systems cannot reliably perform in clinical settings. The task involves dynamic human interaction and judgment that AI cannot execute end-to-end today.
Task automatabilityclaude-sonnet-51/5This requires direct physical manipulation of a patient's body, applying immobilization devices, and adjusting positioning based on real-time physical feedback—a physical manual task current AI cannot perform without embodiment.
Adoption barriersclaude-haiku-4-5-202510015/5Patient handling and positioning is subject to strict occupational safety regulations, healthcare facility protocols, and liability requirements. Direct human contact and judgment in patient care is legally and organizationally protected; automation would face substantial regulatory and institutional resistance.
Adoption barriersclaude-sonnet-55/5Patient handling and safety in medical imaging requires a licensed technologist, involves direct physical contact, liability for patient safety/injury, and regulatory requirements mandating human involvement.
Cost vs. human wageclaude-haiku-4-5-202510011/5The capital cost of a robotic system capable of safe patient handling, combined with integration, maintenance, and required oversight in a hospital setting, far exceeds the labor cost of an MRI technologist performing this task directly.
Cost vs. human wageclaude-sonnet-51/5There is no AI system performing this physical task, so no cost comparison favors AI; a human technologist is required at standard labor cost.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed product performs autonomous patient positioning and immobilization in MRI environments. This remains a physical task requiring embodied robots in sterile, high-stakes medical settings where liability and safety requirements are severe.
Technical feasibility todayclaude-sonnet-51/5No deployed robotic or AI product physically positions and immobilizes patients for MRI scans; this remains entirely a human physical task in clinical practice.

Intravenously inject contrast dyes, such as gadolinium contrast, in accordance with scope of practice.

0

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

Panel ratings per dimension for this task
DimensionModelRatingRationale
Sector adoption velocityclaude-haiku-4-5-202510011/5Healthcare imaging remains heavily dependent on licensed human technologists; adoption of autonomous injection robotics in clinical MRI settings is negligible, with no meaningful production deployment data available.
Sector adoption velocityclaude-sonnet-51/5Healthcare's physical, hands-on procedural tasks show minimal AI adoption; imaging technologist injection duties remain untouched by automation in practice.
Augmentation potentialclaude-haiku-4-5-202510012/5AI could provide minor assistance through real-time image guidance or patient data review to inform injection timing and dosing, but the manual act of injection itself remains human-performed, limiting augmentation value.
Augmentation potentialclaude-sonnet-52/5AI may assist with dosage calculation, contrast reaction risk flagging, or documentation, but offers little to no direct assistance with the physical injection act itself.
Task automatabilityclaude-haiku-4-5-202510011/5Intravenous injection of contrast dye requires precise manual dexterity, real-time patient assessment, and physical contact with the patient's body—capabilities that current AI systems cannot perform. No end-to-end automation exists for this invasive medical procedure.
Task automatabilityclaude-sonnet-51/5This is a hands-on invasive medical procedure requiring physical manipulation, vein assessment, and real-time patient safety monitoring that current AI cannot perform end-to-end.
Adoption barriersclaude-haiku-4-5-202510015/5This task is protected by legal requirements: only licensed MRI technologists or nurses can administer intravenous contrast dyes under scope of practice. Federal and state regulations, liability concerns, and patient safety standards create hard barriers to substitution.
Adoption barriersclaude-sonnet-55/5This is a licensed medical procedure with strict scope-of-practice regulations, direct patient contact, and significant liability for adverse reactions, requiring certified human practitioners.
Cost vs. human wageclaude-haiku-4-5-202510011/5The cost of developing, deploying, and maintaining a robotic system capable of safe IV injection, combined with liability and regulatory oversight, vastly exceeds the cost of a trained technologist performing the task.
Cost vs. human wageclaude-sonnet-51/5There is no viable AI substitute for physical injection, so the comparison is moot and the human-performed task remains the only cost-effective option.
Technical feasibility todayclaude-haiku-4-5-202510011/5No deployed AI product performs intravenous injection autonomously in clinical settings. This task requires embodied robotics with FDA clearance and clinical validation, which does not exist in production at scale for contrast injection.
Technical feasibility todayclaude-sonnet-51/5No deployed product exists that autonomously performs IV contrast injection; this remains firmly a manual clinical task performed by licensed 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.