Nuclear Technicians
19-4051.00Median wage $110,240/yr6,470 employed (US)Rank #901 of 923 scored · top 98% by substitution
Assist nuclear physicists, nuclear engineers, or other scientists in laboratory, power generation, or electricity production activities. May operate, maintain, or provide quality control for nuclear testing and research equipment. May monitor radiation.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution8
Exposure10
Augmentation42
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
18 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
0%
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%11
panel mean rating 1.4/5 → substitution pressure 11/100
Technical feasibility todayw 20%9
panel mean rating 1.4/5 → substitution pressure 9/100
Cost vs. human wagew 15%11
panel mean rating 1.4/5 → substitution pressure 11/100
Adoption barriersw 20%inverted — strong barriers lower the score4
panel mean rating 4.8/5 (barrier strength) → substitution pressure 4/100
Sector adoption velocityw 10%3
panel mean rating 1.1/5 → substitution pressure 3/100
Task breakdown (18 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.
Calculate equipment operating factors, such as radiation times, dosages, temperatures, gamma intensities, or pressures, using standard formulas and conversion tables.
38CI 21–55 · exposure 45 · augmentation 75 · importance 4.0/5 · click for rater detail
Calculate equipment operating factors, such as radiation times, dosages, temperatures, gamma intensities, or pressures, using standard formulas and conversion tables.
38| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities operate under extreme regulatory and safety constraints, with long capital cycles and conservative change management. Adoption of unsanctioned automation in safety-critical calculation tasks is negligible; even pilots are rare because regulatory frameworks require human accountability. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Nuclear industry is a highly regulated, safety-conservative sector with slow technology adoption cycles, though calculation tools have long been used to support technicians. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by auto-populating formulas, flagging unit conversions, or highlighting results that fall outside typical ranges, reducing manual entry errors and increasing review speed. However, the technician must remain fully in control of formula selection and validation, limiting the scope of productivity gain. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI and specialized software strongly augment technicians by rapidly performing standard calculations, reducing error and time while the technician retains oversight and interpretive responsibility. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While the calculation itself (applying formulas to numeric inputs) is automatable, the task requires accurate interpretation of equipment data, selection of correct formulas for context-specific conditions, and validation against safety thresholds. Current AI can perform simple formula application but lacks the domain expertise and physical world grounding to reliably select appropriate formulas and verify that outputs are physically sound in a nuclear safety context. |
| Task automatability | claude-sonnet-5 | 4/5 | These are formulaic, standardized calculations using known conversion tables, well within the capability of AI systems given proper inputs and validated formulas., though safety-critical verification is still needed. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities are among the most heavily regulated industries, with strict licensing and quality assurance requirements. Calculations affecting radiation safety, dosages, and equipment operations must be performed or signed off by licensed technicians; automated systems cannot legally assume final responsibility for these safety-critical outputs without explicit regulatory approval, which does not yet exist. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Nuclear facilities are heavily regulated (NRC oversight), requiring certified technicians and documented calculation procedures with human accountability for safety-critical operations. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | The marginal cost of AI inference for calculation is very low, but integration into legacy nuclear facility systems, validation, and compliance oversight adds material cost. When amortized across typical technician workload and factoring in required human review, costs approach parity with a technician performing the task directly. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated calculation via software/AI is extremely cheap per computation compared to technician time, though initial validation and integration costs exist. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs this task end-to-end in nuclear facilities. Spreadsheet tools and specialized dosimetry software handle rote calculations, but these are purpose-built, not general AI systems. AI would need to extract data from diverse sensor and equipment formats, then validate results—tasks for which production AI systems lack sufficient nuclear domain knowledge and error tolerance. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Calculation tools and software exist for radiation physics and reactor operations, but full deployment requires domain-specific validated systems rather than general-purpose AI, and human verification remains standard practice. |
Monitor nuclear reactor equipment performance to identify operational inefficiencies, hazards, or needs for maintenance or repair.
21CI 18–25 · exposure 30 · augmentation 75 · importance 4.7/5 · click for rater detail
Monitor nuclear reactor equipment performance to identify operational inefficiencies, hazards, or needs for maintenance or repair.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Nuclear power is a highly regulated, risk-averse sector with slow technology adoption cycles. While some facilities pilot AI-assisted monitoring, widespread production deployment remains limited due to regulatory scrutiny, validation requirements, and conservative organizational culture in the nuclear industry. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | The nuclear power sector is highly regulated, safety-conservative, and slow to adopt new technologies, especially for safety-critical monitoring functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI systems demonstrably assist technicians by automating routine data aggregation, flagging sensor anomalies in real-time, and surfacing maintenance recommendations, significantly raising technician productivity while the human remains responsible for judgment and decision-making on critical safety matters. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based anomaly detection, predictive analytics, and pattern recognition can meaningfully assist technicians in identifying trends and potential issues, improving efficiency while humans retain final judgment and authority. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can process sensor data and flag anomalies in real-time, nuclear reactor monitoring requires human judgment, contextual reasoning, and integrated assessment of complex interdependent systems. Current systems can automate alert generation but cannot reliably replace the full monitoring task without expert human oversight, falling short of the 50% time-saving threshold at equal quality. |
| Task automatability | claude-sonnet-5 | 2/5 | While sensor data analysis and anomaly detection can be partially automated, the physical monitoring, safety-critical judgment, and integration with plant control systems require substantial human oversight that AI cannot fully replace end-to-end today. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear reactor operation is heavily regulated (NRC, international standards); licensing laws require trained, certified human operators to monitor and sign off on reactor status. Automation of safety-critical monitoring faces hard legal and liability barriers—a licensed technician must legally remain responsible for reactor safety oversight. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear reactor monitoring is subject to strict NRC (or equivalent) regulations requiring licensed personnel, extensive certification, and legal accountability, making autonomous AI substitution essentially prohibited. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Initial AI system deployment for reactor monitoring is capital-intensive (integration, validation, regulatory certification), and must be combined with human technician oversight. The cost per task-equivalent remains comparable to or higher than human monitoring alone when all implementation and liability factors are included. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Nuclear facilities require redundant, certified, highly reliable monitoring infrastructure; AI systems for this domain require extensive validation, certification, and integration costs that offset savings versus technician wages. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | AI-based monitoring systems exist in nuclear facilities (predictive maintenance, anomaly detection on telemetry), but they operate as decision-support tools with human operators validating all critical alerts. No deployed system autonomously performs end-to-end reactor monitoring without human review and sign-off. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some predictive maintenance and anomaly-detection software is deployed in industrial settings, but nuclear-specific monitoring systems remain heavily human-supervised with narrow AI tool integration, not autonomous production systems. |
Monitor instruments, gauges, or recording devices under direction of nuclear experimenters.
21CI 18–25 · exposure 25 · augmentation 63 · importance 4.6/5 · click for rater detail
Monitor instruments, gauges, or recording devices under direction of nuclear experimenters.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are among the most heavily regulated and conservative adopters; any monitoring AI would require years of validation and regulatory approval, and the sector has shown minimal production adoption of unsupervised AI in safety-critical roles. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Nuclear and scientific research sectors are conservative, safety-driven, and slow to adopt autonomous AI systems for physical instrument monitoring compared to information-sector benchmarks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist technicians by highlighting anomalies, aggregating data, or predicting instrument drift, but the human remains legally and operationally responsible for interpretation and action in this safety-critical domain. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-based anomaly detection, predictive analytics, and automated data visualization can meaningfully assist technicians in tracking instrument readings and flagging deviations, improving efficiency while humans retain control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While routine instrument monitoring could theoretically be automated via sensor feeds and alerts, nuclear environments demand continuous human vigilance for unexpected anomalies and safety-critical deviations that require real-time interpretation—not a straightforward substitution task. |
| Task automatability | claude-sonnet-5 | 2/5 | While sensor data monitoring can be partially automated with software alarms and data logging, this task requires physical presence, real-time judgment under expert direction, and response to novel experimental conditions that current AI cannot fully replicate end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear regulation (NRC, international standards) requires human technicians to be directly responsible and present for critical monitoring; licensing rules and safety liability make autonomous or unsupervised AI substitution legally prohibited, not merely discouraged. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Nuclear facilities are heavily regulated, requiring licensed/certified personnel for safety-critical monitoring, and liability for radiological incidents creates strong institutional resistance to full automation without human sign-off. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Integration of AI monitoring into nuclear facilities requires extensive validation, redundancy, and liability infrastructure; all-in costs rival or exceed the salary of a technician, particularly when regulatory compliance overhead is factored in. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automated sensor systems can reduce some labor costs, but the specialized safety infrastructure, redundancy, and human oversight required in nuclear settings keep all-in AI costs comparable to or only modestly below skilled technician wages. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI systems reliably perform real-time nuclear instrument monitoring in production today; monitoring is either manual or uses traditional automation (alarms, logging) rather than AI agents interpreting complex instrument arrays under regulatory oversight. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Industrial monitoring and anomaly-detection software exists and is used for data logging and threshold alerts, but no deployed AI product autonomously performs the full scope of expert-directed nuclear experiment monitoring reliably in production. |
Communicate with accelerator maintenance personnel to ensure readiness of support systems, such as vacuum, water cooling, or radio frequency power sources.
14CI 5–23 · exposure 13 · augmentation 38 · importance 3.9/5 · click for rater detail
Communicate with accelerator maintenance personnel to ensure readiness of support systems, such as vacuum, water cooling, or radio frequency power sources.
14| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are among the most regulated and conservative sectors, with slow digitization of safety-critical communication and strong resistance to automating coordination tasks that affect public safety and regulatory compliance. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear technician roles operate in a highly regulated, low-digitization physical environment where AI adoption for safety-critical coordination tasks is minimal to nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist by summarizing maintenance logs, drafting routine status updates, or flagging anomalies in system parameters, improving technician efficiency—but human expertise in interpreting complex equipment interactions and managing real-time readiness remains essential. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could assist with scheduling, logging communications, or flagging maintenance status in tracking systems, but offers little help with the substantive judgment and real-time coordination required. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could draft or schedule communications and summarize system status, the task requires real-time coordination with maintenance personnel and dynamic problem-solving for equipment readiness—inherently requiring human judgment and context-awareness that current AI systems cannot fully automate end-to-end. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time interpersonal coordination, physical system checks, and situational judgment in a specialized technical environment that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Nuclear facilities operate under strict NRC and DOE regulations requiring licensed technicians to certify system readiness; liability for equipment failure or safety hazards creates strong legal barriers to full automation, and human accountability for accelerator support system status is deeply embedded in regulatory requirements. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Nuclear facility operations involve strict regulatory oversight, safety certifications, and liability requirements that mandate qualified human personnel for coordination and sign-off on safety-critical systems. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The task involves safety-critical communication in a regulated environment; oversight costs and the need for human verification and accountability mean that AI-assisted or partial automation does not substantially reduce the loaded cost of a qualified technician. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Human technicians with specialized training are required, and there is no AI substitute performing equivalent verification, so cost comparison favors humans by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably handles this task independently. AI chatbots can draft messages but cannot autonomously verify equipment status, negotiate maintenance priorities, or ensure safety-critical system readiness with the accountability nuclear facilities require. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product coordinates accelerator support-system readiness with maintenance personnel; this remains a highly specialized human communication and verification task. |
Test plant equipment to ensure it is operating properly.
10CI 0–20 · exposure 13 · augmentation 50 · importance 4.6/5 · click for rater detail
Test plant equipment to ensure it is operating properly.
10| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Nuclear is a laggard sector for full task automation due to regulatory conservatism, long asset lifecycles, and safety-first culture. While predictive maintenance pilots exist, substantive displacement of testing tasks by autonomous AI is rare in production nuclear plants. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear power is a highly regulated, low-digitization physical industry with slow technology adoption cycles due to safety and compliance requirements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist technicians via predictive analytics, anomaly alerting, and data interpretation dashboards, improving oversight efficiency. However, augmentation is limited by the need for human validation and the regulatory requirement that human technicians retain authority over test decisions. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-based predictive maintenance and anomaly detection software can assist technicians by flagging irregularities in sensor data, improving efficiency of the testing process without replacing the hands-on task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Testing nuclear plant equipment requires physical inspection, sensor deployment, and real-time judgment in safety-critical environments. While AI can analyze historical data and predict anomalies, the core task of in-situ equipment testing and ensuring proper operation involves hands-on intervention that current autonomous systems cannot reliably perform in complex, hazardous nuclear settings. |
| Task automatability | claude-sonnet-5 | 1/5 | Physical inspection and functional testing of nuclear plant equipment requires hands-on sensor readings, physical manipulation, and situational judgment that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities are heavily regulated (NRC, IAEA) with strict licensing requirements. Federal law mandates that licensed operators and technicians perform safety-critical tests; automation of testing is tightly constrained by regulatory mandate and liability asymmetry around failures. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear facilities are heavily regulated (NRC oversight), requiring licensed technicians for safety-critical equipment testing, with strict liability and legal authorization requirements. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-based monitoring tools cost significant capital and ongoing integration/oversight. Trained nuclear technicians earn $70k–$100k+ loaded, and AI does not yet replace their full testing function; systems remain supplementary and not cheaper all-in. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | Physical testing requires human presence, specialized instruments, and regulatory sign-off, making AI substitution infeasible and not cost-competitive for the physical task itself. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed AI system reliably performs end-to-end nuclear equipment testing. Monitoring and anomaly detection exist, but actual plant testing—sensor placement, calibration, validation under nuclear conditions—remains human-dependent. Research systems exist, but production deployment in nuclear facilities is minimal. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product autonomously performs physical equipment testing in nuclear plants; this remains firmly a human/technician task with sensor-based monitoring as an aid at best. |
Set up equipment that automatically detects area radiation deviations and test detection equipment to ensure its accuracy.
9CI 0–18 · exposure 13 · augmentation 38 · click for rater detail
Set up equipment that automatically detects area radiation deviations and test detection equipment to ensure its accuracy.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are highly regulated, risk-averse environments with strict personnel qualification requirements. Adoption of AI-driven automation for safety-critical radiation detection setup is minimal; the sector moves slowly due to regulatory scrutiny and safety culture that privileges human expertise and accountability. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | The nuclear power sector is a highly regulated, low-digitization physical environment with minimal AI agent deployment for hands-on equipment tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist technicians by automating data logging, flagging anomalies in test results, and generating compliance reports, improving their efficiency. However, the core tasks of physical equipment setup and validation testing remain human-centric, limiting augmentation to supporting workflows rather than transforming core performance. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could assist with logging calibration data, generating maintenance schedules, or analyzing test results, but offers little help with the physical setup and testing itself. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist with data analysis of radiation readings and alert thresholds, the physical setup of detection equipment and hands-on testing require human technicians to handle hardware, calibrate sensors, and perform field measurements. No current system can reliably perform the full end-to-end task of equipment setup and validation testing without substantial human intervention. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical setup and calibration of radiation detection hardware in a controlled nuclear facility, which current AI systems cannot physically perform. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | This task is heavily regulated by the Nuclear Regulatory Commission (NRC) and similar bodies worldwide. Licensed nuclear technicians must legally set up and certify detection equipment; regulatory frameworks explicitly require qualified human personnel to sign off on radiation safety equipment. Liability and safety requirements create hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear facilities operate under strict NRC/regulatory licensing requiring certified technicians to perform and verify safety-critical radiation detection equipment tasks. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Nuclear technicians command high wages due to licensing and training requirements. AI systems that assist with monitoring are relatively expensive to deploy and maintain in safety-critical settings, and still require paid technician oversight, making the total cost comparable to or higher than human-only approaches. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical labor and certified expertise involved, so there is no viable cost comparison favoring AI today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some software-based monitoring and alert systems exist to process radiation data, but no deployed AI product reliably performs the complete task of setting up and physically testing detection equipment in real-world nuclear environments. Most solutions are narrow (data analysis only) and require human technician oversight throughout setup and validation. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical installation and calibration testing of radiation monitoring equipment; this remains a hands-on technician task. |
Conduct surveillance testing to determine safety of nuclear equipment.
9CI 0–18 · exposure 13 · augmentation 50 · importance 4.8/5 · click for rater detail
Conduct surveillance testing to determine safety of nuclear equipment.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear power is a highly regulated, risk-averse sector with slow technology adoption cycles; facilities are reluctant to replace certified human inspectors with automated systems due to safety criticality and regulatory conservatism. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | The nuclear power sector is a highly regulated, low-digitization, physical-infrastructure environment with minimal AI agent deployment for safety-critical testing. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI systems can assist technicians by highlighting sensor anomalies, organizing inspection data, and flagging trends, usefully supporting human decision-making, though the core safety determination remains human-centric. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with data logging, anomaly detection in sensor readings, and predictive maintenance analytics to support technicians, though the core testing remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI could assist in data collection and anomaly detection from sensor readings, the task requires physical inspection, judgment about equipment safety, and regulatory compliance verification that demands human expertise and physical presence on-site. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical inspection, hands-on operation of test equipment, and real-time judgment in a nuclear facility environment that current AI cannot perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facility operations are heavily regulated by the NRC and other bodies; qualified nuclear technicians must conduct safety certifications, and liability for equipment failures is severe, creating hard legal and regulatory barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear safety testing is governed by strict NRC licensing and regulatory requirements mandating qualified human technicians to perform and certify these tests. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-based monitoring systems are expensive to deploy and integrate into nuclear facilities, and they require extensive human oversight and verification, making the all-in cost comparable to or exceeding trained human technician labor. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the certified human labor and physical presence required, so no cost-per-task-equivalent comparison favors AI. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Some automated monitoring systems exist for continuous surveillance, but no deployed product reliably performs the full safety determination end-to-end; human technicians must interpret complex equipment states and certify safety in real nuclear facilities. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product autonomously conducts nuclear surveillance testing; this remains a highly regulated, human-performed physical and procedural task. |
Measure the intensity and identify the types of radiation in work areas, equipment, or materials, using radiation detectors or other instruments.
9CI 0–18 · exposure 13 · augmentation 38 · importance 4.0/5 · click for rater detail
Measure the intensity and identify the types of radiation in work areas, equipment, or materials, using radiation detectors or other instruments.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear technician work remains highly regulated and centralized in established facilities with conservative adoption practices. Digital transformation in this sector is slow; most organizations maintain human-in-the-loop protocols for radiation measurement due to safety and compliance requirements. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | The nuclear industry is highly regulated, safety-conservative, and slow to adopt automation for core safety-critical physical tasks, with radiation surveying remaining manual and human-performed. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by automating data interpretation from detector readings, suggesting radiation type classification, or flagging anomalies for technician review. However, the core task of measurement and on-site decision-making remains human-dependent, limiting augmentation scope. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can help analyze historical radiation data, predict contamination patterns, or assist in reporting and documentation, but it offers little direct assistance to the physical act of measuring radiation in real time. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can analyze radiation detector output data and classify types of radiation from spectroscopic readings, the physical act of positioning detectors in work areas, interpreting real-time environmental context, and making safety-critical decisions requires human presence and judgment. Automation would cover only a narrow portion of the full task. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical presence with handheld or fixed radiation detection instruments to survey actual work areas, equipment, and materials in a nuclear facility; no AI system can physically operate a Geiger counter or dosimeter in a plant.The task is fundamentally a physical measurement action, not information processing. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities operate under strict regulatory oversight (NRC, IAEA); radiation safety work typically requires licensed professionals with legal accountability for measurements and safety determinations. Autonomous measurement without licensed human sign-off faces substantial regulatory and liability barriers. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear safety regulations (NRC and similar bodies) require certified/licensed personnel to perform radiation monitoring and hold accountability for facility safety, making this a hard-barrier task requiring human authorization and liability assumption. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Radiation measurement requires specialized hardware and often trained personnel on-site; AI systems would require integration with existing detection equipment, calibration, and continuous oversight. Total cost would remain comparable to or exceed a technician's time for most real-world deployments. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute for the physical sensing and mobility required, so cost comparison favors the human technician who must physically operate detection equipment; robotic radiation surveying exists in niche cases but is far more costly than routine human measurement. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Detection instruments exist but are primarily traditional hardware (Geiger counters, spectrometers); limited autonomous systems perform end-to-end measurement and classification without human technician intervention. AI-enhanced analysis of detector data is emerging but not yet reliably deployed at scale in production nuclear environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical radiation surveys; this remains a hands-on task performed by certified technicians using calibrated instruments, with AI at most assisting in data logging afterward. |
Decontaminate objects by cleaning them using soap or solvents or by abrading using brushes, buffing machines, or sandblasting machines.
9CI 0–18 · exposure 8 · augmentation 25 · importance 3.7/5 · click for rater detail
Decontaminate objects by cleaning them using soap or solvents or by abrading using brushes, buffing machines, or sandblasting machines.
9| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities operate in a highly regulated, conservative sector with strong worker representation and slow technology adoption cycles. Decontamination remains largely manual despite decades of robotics development, reflecting deep institutional and regulatory resistance to automation in this domain. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear technician work is a highly specialized, low-digitization physical trade with minimal AI adoption; the sector shows no meaningful trend toward AI-driven automation of decontamination procedures. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist technicians by predicting optimal decontamination methods based on contamination type and material, or by monitoring robotic systems and flagging anomalies, but current tools provide only partial assistance on specific sub-tasks rather than transformative productivity gains. |
| Augmentation potential | claude-sonnet-5 | 1/5 | AI offers essentially no assistance to the physical act of scrubbing, sandblasting, or solvent-cleaning contaminated objects, though it could tangentially help with logging or scheduling unrelated to the core physical task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While physical cleaning equipment like brushes and sandblasting machines could theoretically be automated, this task requires judgment about appropriate decontamination methods, material compatibility, and verification of cleanliness—decisions that current AI systems cannot reliably make autonomously in nuclear safety contexts. Automation of the physical motion alone would save <50% of total task time given the need for human oversight and decision-making. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, hands-on decontamination task involving direct manipulation of contaminated objects with cleaning tools and abrasive equipment; current AI systems cannot perform physical manipulation tasks like this. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear decontamination is heavily regulated by the NRC and DOE; human technicians must oversee and validate all decontamination work, and regulatory frameworks require human authorization and responsibility for decontamination outcomes. Liability and safety certification create hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Nuclear decontamination work is subject to strict radiation safety regulations, certification requirements, and specialized handling protocols that require trained, often licensed personnel, creating strong regulatory and safety barriers to substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Robotic decontamination systems capable of nuclear-grade reliability are extremely expensive to acquire, maintain, and integrate, likely exceeding the cost of trained nuclear technicians. Current systems require extensive setup and oversight per-site. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI software has no direct cost application here since the task requires physical robotic hardware and specialized equipment, not inference-based automation; any robotic solution would be far more costly than trained technician labor for a niche task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI systems perform autonomous nuclear decontamination. This task requires robots with real-time material assessment, decision-making under contamination uncertainty, and adherence to strict nuclear safety protocols—capabilities not currently available in production systems at scale. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical decontamination scrubbing, sandblasting, or solvent cleaning of nuclear materials; this remains entirely a manual/robotic-mechanical task outside AI's software domain. |
Apply safety tags to equipment needing maintenance.
4CI 0–9 · exposure 8 · augmentation 25 · importance 4.6/5 · click for rater detail
Apply safety tags to equipment needing maintenance.
4| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are among the most conservative and heavily regulated sectors with minimal AI adoption in safety-critical physical tasks; regulatory approval and safety certification timelines are extremely long. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear power is a highly regulated, safety-critical, low-digitization physical sector with minimal AI adoption for hands-on safety procedures. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist with identifying which equipment requires maintenance through data analysis of maintenance schedules, but the physical application of tags and safety verification inherently requires human presence. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI could help track maintenance schedules or generate tagging checklists digitally, but offers little assistance to the core physical act of applying tags. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Applying physical safety tags requires robotic manipulation in specific spatial contexts and decision-making about which equipment needs tagging—capabilities that are not yet reliable or cost-effective in nuclear facilities where precision and safety are critical. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, hands-on task requiring a technician to be present at equipment locations to attach physical tags, which current AI systems cannot perform without robotic embodiment.dd |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities are heavily regulated with strict licensing requirements; safety-critical tasks like equipment tagging must be performed or verified by licensed nuclear technicians, creating a hard legal barrier to full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear safety tagging is governed by strict regulatory lockout/tagout and NRC procedures requiring qualified, authorized personnel to physically verify and tag equipment, with high liability for errors. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current robotic systems capable of precise physical manipulation and integration into nuclear facility workflows would cost significantly more than the technician labor required. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this physical action, so AI cost cannot be compared favorably; humans remain the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed commercial product reliably performs end-to-end physical tagging in nuclear environments; this remains research-stage for robotics and would require specialized integration. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product physically applies lockout/tagout safety tags to equipment in nuclear facilities; this remains entirely a manual human task. |
Collect air, water, gas or solid samples for testing to determine radioactivity levels or to ensure appropriate radioactive containment.
3CI 0–5 · exposure 0 · augmentation 25 · importance 3.5/5 · click for rater detail
Collect air, water, gas or solid samples for testing to determine radioactivity levels or to ensure appropriate radioactive containment.
3| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities operate under strict legacy protocols and regulatory oversight; adoption of fully autonomous sample collection is negligible in practice. The sector remains highly conservative and human-dependent for safety-critical sampling tasks. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear/utilities sector is a slow-adopting, highly regulated, physical-operations sector with minimal AI-driven automation of field sampling tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI-powered radiation sensors and real-time monitoring dashboards can assist technicians in deciding where and what to sample, but the core act of physically collecting samples remains manual. Limited augmentation potential exists because the bottleneck is the physical act of collection rather than decision-making. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with logging results, scheduling, or analyzing radioactivity data after collection, but offers little help with the physical act of sample collection itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Sample collection in nuclear environments requires physical presence, real-time hazard assessment, and adaptation to containment protocols that current AI cannot perform autonomously. The task demands situated judgment about radiation safety and proper sampling technique that cannot be remotely or fully automated today. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical sampling task requiring on-site presence, manipulation of equipment in radiologically controlled areas, and adherence to chain-of-custody protocols; no AI system can physically collect samples. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear technician work is heavily regulated by the Nuclear Regulatory Commission (NRC) and other bodies; sampling procedures require licensed or certified human technicians to personally perform and document chain-of-custody. Legal and regulatory requirements mandate human accountability and physical presence for radioactive sample handling. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Nuclear facilities operate under strict regulatory oversight (NRC and similar bodies) requiring trained, often certified personnel to handle radioactive sampling, with significant liability and safety protocols limiting substitution. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot substitute for the physical collection task itself, so the cost ratio is dominated by the human technician performing the work. Automation would require robotic systems and remote handling equipment, which add significant capital cost rather than reducing per-task labor cost. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for physical sample collection, so AI cost is irrelevant/infinite relative to a human technician performing the task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can autonomously collect physical samples from air, water, gas, or solid media in nuclear facilities. While sensors can measure radioactivity, the human task of sample acquisition remains entirely manual and human-dependent in production nuclear operations. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI or robotic product performs autonomous environmental radiological sampling in nuclear facilities today; this remains a manual, specialized task. |
Determine or recommend radioactive decontamination procedures, according to the size and nature of equipment and the degree of contamination.
0CI 0–0 · exposure 0 · augmentation 38 · click for rater detail
Determine or recommend radioactive decontamination procedures, according to the size and nature of equipment and the degree of contamination.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear energy remains a laggard sector for AI adoption due to strict licensing requirements, safety culture, and regulatory barriers. Decontamination procedures are not candidates for autonomous AI deployment. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear industry is highly regulated, safety-conservative, and slow to adopt automation for core safety-critical technical judgments. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist a human technician by retrieving regulatory guidance, equipment manuals, or historical decontamination protocols from a database, but the task's safety-critical judgment and regulatory sign-off requirements limit meaningful augmentation. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by referencing decontamination protocols, historical data, and documentation support, but the core recommendation still relies on human expertise and on-site assessment. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires synthesizing specialized knowledge of radiological hazards, equipment-specific decontamination protocols, regulatory compliance, and real-time assessment of contamination severity. Current AI lacks the domain expertise, physical inspection capability, and liability-bearing judgment to make independent decontamination recommendations in safety-critical nuclear environments. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical assessment of contaminated equipment, real-time radiation measurement, and safety-critical judgment that current AI cannot perform end-to-end without extensive human sensing and decision-making. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear decontamination is heavily regulated by the NRC, EPA, and other agencies. A licensed nuclear technician or qualified health physicist must legally perform or directly approve decontamination procedures; automation is prohibited by regulatory requirement and liability law. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear safety regulations mandate qualified, often licensed personnel to make contamination and decontamination decisions, with severe liability and regulatory oversight (e.g., NRC) preventing AI autonomy. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of an AI system capable of handling this task (training, regulatory validation, integration, oversight) would far exceed the cost of a trained nuclear technician, given the low volume and high liability of decontamination work. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot independently perform the physical inspection and regulatory-compliant decision-making required, so there is no meaningful cost substitution possible today. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably makes radioactive decontamination procedure recommendations in production. This requires integration with dosimetry data, regulatory frameworks (NRC, IAEA), and equipment-specific technical documentation—maturity is research-stage only. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI products perform decontamination procedure determination in nuclear facilities; this remains a specialized human technical judgment task governed by safety protocols. |
Follow nuclear equipment operational policies and procedures that ensure environmental safety.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.8/5 · click for rater detail
Follow nuclear equipment operational policies and procedures that ensure environmental safety.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear is a laggard sector for autonomous AI adoption in safety-critical tasks. Regulatory conservatism, long facility lifespans, and the catastrophic-failure paradigm mean deployment velocity for AI in autonomous nuclear safety roles is near-zero and will remain so. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | The nuclear power sector is slow to adopt AI for safety-critical operational roles due to regulatory conservatism and high stakes of error. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI could assist technicians by automating routine compliance documentation, flagging anomalies in sensor data, or providing rapid access to procedure references. However, the core safety-decision loop remains human-centered, limiting augmentation to supporting rather than transforming productivity. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with monitoring dashboards, anomaly detection, or documentation support, but the core procedural compliance and safety judgment remains human-driven with limited AI augmentation. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Following nuclear safety procedures requires continuous judgment, real-time environmental monitoring, and adaptive decision-making under high stakes. While AI could assist in documentation, the core responsibility—ensuring safety compliance through situational reasoning and discretionary application of procedures—remains fundamentally human. |
| Task automatability | claude-sonnet-5 | 1/5 | This task requires physical presence, situational judgment, and adherence to safety-critical procedures in a nuclear facility, which current AI cannot execute end-to-end.atorio. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities are among the most heavily regulated industries; NRC (and international equivalents) mandate that licensed nuclear operators and technicians must personally perform and sign off on safety-critical procedures. Legal and liability barriers are absolute—automation of core safety functions is prohibited by design. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear operations are heavily regulated (NRC licensing, mandatory certified personnel, strict liability), making this one of the most protected task categories against automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A nuclear technician's loaded cost is high (~$70–90k/year including benefits and training), but the cost of AI failures in nuclear environments—regulatory penalties, safety incidents, liability—far exceeds the human wage. AI deployment would require extensive redundant oversight that eliminates any cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this physical, compliance-driven task, so cost comparison favors the human by default. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system operates nuclear equipment or makes real-time safety decisions autonomously in production nuclear facilities. Regulatory frameworks and safety-critical protocols require qualified human technicians; no commercial product performs independent nuclear safety oversight. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product autonomously follows nuclear operational safety procedures on the plant floor; this remains a human-executed, regulated function. |
Follow policies and procedures for radiation workers to ensure personnel safety.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.6/5 · click for rater detail
Follow policies and procedures for radiation workers to ensure personnel safety.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facility operations are highly conservative, tightly regulated, and deeply embedded in safety-critical human oversight; adoption of autonomous AI for safety enforcement is not occurring and faces regulatory prohibition. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear facility operations are a highly regulated, physical, low-digitization environment where AI adoption for hands-on safety compliance is minimal to nonexistent. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with data logging, compliance record-keeping, or flagging anomalies in radiation exposure data, but the core task of ensuring personnel safety through judgment and enforcement requires sustained human presence and accountability. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled monitoring systems, dosimetry alerts, and digital checklists can help technicians track and remember procedural steps, offering moderate assistance without performing the compliance task itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task is inherently supervisory and requires real-time judgment about worker compliance, environmental conditions, and safety protocol interpretation in live settings. AI cannot independently monitor actual personnel behavior or make dynamic safety decisions that depend on physical context and human accountability. |
| Task automatability | claude-sonnet-5 | 1/5 | This task is fundamentally about personal compliance behavior in a physical, high-hazard environment—donning dosimeters, following exclusion zones, executing procedures correctly on the job floor—which requires embodied human action, not information processing that AI can substitute for. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities are among the most heavily regulated industries; radiation safety officers must be licensed professionals who are legally responsible for worker safety, and regulatory bodies (NRC, IAEA) mandate human accountability for radiation protection decisions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear regulatory frameworks (e.g., NRC requirements) mandate that certified personnel physically follow radiation safety protocols, and liability for radiation exposure incidents is severe, making human execution and accountability legally required. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A human radiation safety officer (with required credentials and liability insurance) costs far less than the combination of comprehensive monitoring AI, integration into nuclear facilities, continuous oversight, and regulatory compliance infrastructure needed to replace this role. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no AI substitute performing this task, so no meaningful cost comparison exists; any AI cost would be additive to, not a replacement for, the human labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can reliably oversee radiation worker safety compliance or make binding safety decisions in nuclear environments; this remains a human responsibility with legal and liability requirements that preclude autonomous AI deployment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs the physical act of following radiation safety procedures; at most, software can log compliance or send reminders, but the task itself remains human-executed. |
Modify, devise, or maintain nuclear equipment used in operations.
0CI 0–0 · exposure 0 · augmentation 25 · importance 4.4/5 · click for rater detail
Modify, devise, or maintain nuclear equipment used in operations.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities operate under strict regulatory oversight that prohibits autonomous systems from performing critical equipment maintenance or modifications; adoption of AI for this specific task is effectively zero across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | The nuclear power/energy sector is a highly regulated, low-digitization, physical-infrastructure sector with slow AI adoption for hands-on equipment work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might provide limited assistance through diagnostics, documentation retrieval, or procedure guidance, but the core physical and judgment-intensive work remains human-dependent and current assistive applications are minimal in nuclear operations. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI can assist with diagnostics, predictive maintenance scheduling, or documentation, but offers limited direct assistance for the physical modification and hands-on maintenance work itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Nuclear equipment modification and maintenance requires hands-on physical manipulation, real-time environmental assessment, safety protocol compliance, and high-stakes decision-making in radioactive environments. Current AI systems cannot perform the physical work, assess radiation hazards directly, or make the nuanced safety judgments required. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a hands-on physical task requiring physical manipulation, fabrication, and modification of specialized nuclear equipment, which current AI cannot perform end-to-end.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear regulation (NRC, international atomic energy standards) mandates that licensed nuclear technicians perform or directly supervise equipment modifications and maintenance. Legal liability and safety protocols create hard barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear equipment maintenance is heavily regulated (e.g., NRC oversight), requires licensed/certified technicians, and carries severe safety and liability consequences, creating hard legal and physical barriers to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems cannot physically access or modify nuclear equipment, and the specialized labor (licensed nuclear technicians earning significant wages) would still be required for legal and safety reasons, making automation economically irrelevant. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for the physical labor and specialized manual work involved, so AI cannot reduce cost relative to the human technician doing this task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product performs end-to-end nuclear equipment modification or maintenance in operational nuclear facilities. This task requires human presence in restricted nuclear environments and licensed technician judgment. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed products autonomously modify or maintain nuclear equipment; this remains firmly in the domain of trained human technicians with physical dexterity and specialized certification. |
Perform testing, maintenance, repair, or upgrading of accelerator systems.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.3/5 · click for rater detail
Perform testing, maintenance, repair, or upgrading of accelerator systems.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are heavily regulated, conservative in adopting new technologies, and require human licensure for safety-critical work. Adoption of AI automation for accelerator maintenance is negligible and unlikely to accelerate given regulatory constraints. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear and specialized physical infrastructure sectors show minimal AI adoption for hands-on technical repair work, reflecting slow digitization of physical maintenance tasks. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist with diagnostics or documentation review, but the hands-on, safety-critical nature of the work and need for real-time judgment in specialized equipment limits meaningful augmentation. Current tools offer minimal productivity gains for core maintenance tasks. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist with diagnostics, predictive maintenance analytics, and documentation support, improving technician efficiency without replacing the physical repair work itself. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Accelerator systems maintenance requires hands-on physical work, real-time diagnostics, safety-critical decision-making, and domain expertise that current AI cannot perform end-to-end. The task involves complex equipment troubleshooting and repair that depends on embodied presence and judgment in highly specialized environments. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires hands-on physical inspection, repair, and calibration of complex accelerator hardware, which current AI cannot perform end-to-end without robotic embodiment far beyond today's capabilities.rating remains low as most of the physical labor cannot be automated. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear facilities operate under strict NRC and international regulatory oversight requiring licensed personnel to perform maintenance and testing. Legal liability, safety certification requirements, and mandatory human sign-off on critical systems create hard legal barriers to automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear facility work involves strict regulatory oversight, safety certifications, and licensing requirements that mandate qualified human technicians for hands-on system work. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | AI systems capable of accelerator maintenance would require custom integration, specialized sensors, robotic platforms, and extensive validation—costs far exceeding the salary of skilled nuclear technicians. Current general-purpose AI adds no cost advantage for this task. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI systems cannot substitute for the skilled labor and physical intervention required, so there is no cost basis for comparison; humans remain the only viable option. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI system can reliably perform testing, maintenance, repair, or upgrading of accelerator systems independently. This requires specialized robotics, domain knowledge, and safety protocols that are not mature in production environments for this specific application. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously tests, maintains, or repairs particle accelerator systems; this remains a highly specialized human technical function. |
Warn maintenance workers of radiation hazards and direct workers to vacate hazardous areas.
0CI 0–0 · exposure 0 · augmentation 38 · importance 4.1/5 · click for rater detail
Warn maintenance workers of radiation hazards and direct workers to vacate hazardous areas.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are among the most heavily regulated sectors; adoption of autonomous AI for safety-critical radiation warnings is essentially non-existent and legally prohibited in practice. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear power and related industries are highly regulated, safety-conservative, and slow to adopt autonomous systems for hazard control functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI might assist by providing real-time radiation monitoring data or alerting summaries to a technician, but the core task of warning and directing workers remains human-dependent due to safety and regulatory requirements. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-enabled sensors, dosimetry alerts, and monitoring dashboards can help technicians detect and communicate hazards faster, but the human still directs evacuation and makes final calls. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time perception of worker locations, radiation field measurement, dynamic hazard assessment, and human-directed communication in a safety-critical environment. Current AI systems cannot reliably detect, localize, and direct human behavior in nuclear facilities with the precision and accountability this task demands. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires real-time physical presence, judgment about ambiguous hazard conditions, and direct human authority to compel evacuation—AI cannot perform this physical safety-critical task end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear Regulatory Commission (NRC) regulations, worker safety laws, and industry standards require a qualified human technician to identify hazards and direct evacuations. Liability for radiation exposure makes this a licensed, legally mandated human responsibility. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear safety regulations (NRC and similar bodies) mandate qualified, licensed personnel to monitor radiation and direct hazard response, with severe liability for failures, making this a hard-barrier task. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The infrastructure, sensors, integration, and continuous oversight required to deploy an AI system for radiation safety would far exceed the cost of a trained nuclear technician performing this task directly. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute performing this task alone, so any cost comparison favors the human who carries legal responsibility and situational judgment. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs autonomous radiation hazard warnings and worker evacuation direction in operational nuclear facilities. This remains a human responsibility requiring licensed judgment and legal accountability in a heavily regulated domain. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed product autonomously monitors radiation and directs human workers to evacuate in real nuclear facility operations; this remains a human safety function with sensor-assist at most. |
Identify and implement appropriate decontamination procedures, based on equipment and the size, nature, and type of contamination.
0CI 0–0 · exposure 0 · augmentation 38 · importance 3.8/5 · click for rater detail
Identify and implement appropriate decontamination procedures, based on equipment and the size, nature, and type of contamination.
0| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Nuclear facilities are highly conservative, heavily regulated sectors with minimal AI adoption for safety-critical tasks. Procedural changes and automation in decontamination face long approval cycles and institutional resistance due to safety and compliance requirements. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Nuclear facilities are a highly regulated, low-digitization physical sector with minimal AI agent deployment for hands-on decontamination work. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by suggesting decontamination protocols based on contamination type or flagging sensor data anomalies, but the safety-critical nature of contamination assessment and the technician's need for real-time judgment limit meaningful augmentation; most value remains with human expertise. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can assist by providing decision-support for identifying contamination type and referencing appropriate protocols, though the physical assessment and execution remain human-driven. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time visual assessment of contamination, judgment of equipment-specific protocols, and safety-critical decision-making that cannot be fully automated. Current AI systems lack the embodied sensing, dynamic environmental reasoning, and accountability for life-safety outcomes needed to independently select and execute decontamination procedures. |
| Task automatability | claude-sonnet-5 | 1/5 | This requires physical assessment of radioactive contamination and hands-on execution of decontamination procedures in hazardous environments, which no current AI system can perform end-to-end. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Nuclear decontamination is heavily regulated by the NRC and other authorities, with strict licensing and certification requirements for personnel performing these tasks. Liability for contamination errors is asymmetric and severe, and regulatory frameworks legally mandate human expertise and sign-off on decontamination decisions. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Nuclear safety regulations, licensing requirements, and liability for radiological hazards mandate qualified human technicians to identify and execute these procedures. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of AI systems with sufficient sensing, robotic capability, and safety integration to handle nuclear decontamination would far exceed the loaded cost of a trained nuclear technician, even accounting for integration and oversight overhead. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | AI cannot substitute for the physical, judgment-intensive, safety-critical work involved, so there is no viable cost comparison for full task replacement. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs end-to-end decontamination procedure selection and implementation in nuclear facilities. This requires physical interaction, real-time hazard assessment, and regulatory compliance verification that exists only in research or highly constrained lab settings, not production nuclear environments. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed products autonomously identify and execute nuclear decontamination procedures; this remains a specialized human technical function governed by safety protocols. |
Related occupations — Life, Physical & Social Science
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.