Role Definition
| Field | Value |
|---|---|
| Job Title | Radiation Protection Adviser (RPA) |
| Seniority Level | Mid-Level (3-7 years post-accreditation) |
| Primary Function | Statutory adviser appointed under Regulation 14 of the Ionising Radiations Regulations 2017 (IRR17). Provides expert radiation protection advice to employers working with ionising radiation. Advises on dose limitation, dose constraint setting, designation of controlled/supervised areas, shielding design for X-ray rooms and sealed source facilities, radioactive waste categorisation and disposal, contingency planning, and personal dosimetry programmes. Works across nuclear sites (AWE, Sellafield, EDF), hospitals (NHS radiology, nuclear medicine, radiotherapy), and industrial radiography (NDT companies, research labs). Must hold RPA2000 accreditation — the UK recognition scheme for RPAs under IRR17. Conducts site inspections, radiation risk assessments, prior risk assessments for new installations, and critical examination of engineering controls. |
| What This Role Is NOT | Not a Health Physics Technician / Radiation Protection Technician (hands-on surveys, dosimetry management — scores 48.5). Not a Radiation Protection Supervisor (employer-appointed operational role under IRR17 Reg 18, not requiring RPA2000). Not a Medical Physics Expert (clinical role under IR(ME)R 2017 providing physics advice on patient dose). Not a Nuclear Engineer (designs reactor systems, fuel cycles — scores 58.6). |
| Typical Experience | 3-7 years post-RPA2000 accreditation. Degree in physics, radiological protection, nuclear engineering, or medical physics. RPA2000 Certificate of Competence in relevant scope (medical, industrial, nuclear, or combination). May hold additional Radioactive Waste Adviser (RWA) or Dangerous Goods Safety Adviser (DGSA) certification. Many RPAs register with SRP (Society for Radiological Protection) at IRadP or CRadP level. UK-specific statutory role — no direct US equivalent (US uses RSO — Radiation Safety Officer under NRC). ~700-1,000 RPA2000-accredited individuals in the UK. |
Seniority note: Trainee RPAs (RPAs in training / RPTs pursuing RPA2000) would score lower — less advisory autonomy, more supervised support work. Senior/Principal RPAs leading radiation protection programmes across multiple sites with corporate advisory responsibility would score higher Green.
Protective Principles + AI Growth Correlation
| Principle | Score (0-3) | Rationale |
|---|---|---|
| Embodied Physicality | 2 | Conducts site inspections of controlled/supervised areas, physically verifies shielding installations, inspects source storage and waste facilities, walks down X-ray rooms and industrial radiography enclosures. Semi-structured environments but includes radiologically controlled areas in hospitals, nuclear facilities, and industrial sites. Not desk-only — the site visit is fundamental to statutory advisory function. |
| Deep Interpersonal Connection | 1 | Briefs employers and RPSs on regulatory requirements, communicates dose investigation findings to workers, advises clinical teams on shielding adequacy, trains staff on local rules and contingency plans. Trust matters — employers rely on RPA judgment for regulatory compliance — but the relationship is professional-advisory, not therapeutic. |
| Goal-Setting & Moral Judgment | 2 | Makes safety-critical advisory judgments — whether proposed shielding is adequate, whether a dose investigation indicates systemic failure, whether waste categorisation meets environmental permit conditions, whether an employer's arrangements satisfy IRR17. Interprets how regulations apply to novel situations (new scanning equipment, unusual isotopes, non-standard facility layouts). Advisory liability — if RPA advice is negligent and results in overexposure, consequences include HSE enforcement, loss of RPA2000 accreditation, and civil/criminal proceedings. |
| Protective Total | 5/9 | |
| AI Growth Correlation | 0 | Demand driven by IRR17 statutory requirement, not AI adoption. Every employer using ionising radiation above exemption levels must appoint an RPA. AI neither creates nor reduces this regulatory mandate. Minor new tasks emerging (validating AI-driven dose monitoring systems) but insufficient for +1. |
Quick screen result: Moderate protective principles (5/9) with strong regulatory barriers. Predicts Green Zone, likely Transforming given mix of automatable calculation/documentation tasks and protected statutory advisory/inspection work.
Task Decomposition (Agentic AI Scoring)
| Task | Time % | Score (1-5) | Weighted | Aug/Disp | Rationale |
|---|---|---|---|---|---|
| Statutory radiation protection advice & dose limit compliance | 20% | 2 | 0.40 | AUG | Core advisory function — advising employers on compliance with IRR17 dose limits (20 mSv/year classified workers, 6 mSv/year others, 1 mSv/year public), dose constraints, and ALARP optimisation. Interpreting how regulations apply to specific workplace situations. AI can present dose data summaries but cannot hold statutory RPA appointment or bear advisory liability. Human judgment required for novel situations and regulatory interpretation. |
| Shielding design assessment & critical examination | 15% | 3 | 0.45 | AUG | Designs and verifies shielding for X-ray rooms, CT suites, PET facilities, sealed source stores, industrial radiography enclosures. AI handles significant sub-workflows — Monte Carlo shielding calculations, workload modelling, parametric dose assessment. But the RPA validates assumptions, inspects physical installations, and certifies adequacy considering real-world usage patterns and occupancy factors that deviate from textbook scenarios. |
| Site inspections & radiation risk assessment | 15% | 1 | 0.15 | NOT | Physically inspects controlled and supervised areas, verifies engineering controls, checks source inventories, assesses workplace arrangements, identifies non-compliance. Walks hospital departments, nuclear facilities, and industrial sites. Observes actual working practices versus documented procedures. Irreducibly physical — must see the facility, talk to workers, check equipment in situ. |
| Radioactive waste categorisation & disposal advice | 10% | 3 | 0.30 | AUG | Advises on waste categorisation (VLLW, LLW, ILW, HLW), waste disposal routes, environmental permit compliance, and waste minimisation. AI handles significant sub-workflows — radionuclide decay calculations, waste accumulation tracking, disposal route optimisation against permit conditions. But categorisation decisions at boundaries require professional judgment, especially for mixed-waste scenarios and unusual isotope combinations. |
| Dose investigation & personal dosimetry programme management | 10% | 3 | 0.30 | AUG | Investigates overexposures, dose anomalies, and lost dosimeter events. Reviews personal dose records, advises on dosimetry requirements, recommends classification of workers. AI-driven dose tracking systems automate routine monitoring and flag exceedances. Investigation of root causes and determination of whether a dose investigation is required under IRR17 Reg 25 requires human judgment. Routine dose record management is increasingly automated. |
| Compliance documentation & local rules | 10% | 4 | 0.40 | DISP | Drafts local rules, writes radiation risk assessments, prepares systems of work, produces compliance reports and audit findings. Structured, template-driven documentation that AI agents can generate end-to-end from regulatory templates and site-specific data. RPA reviews for technical accuracy and site-specific applicability, but the authoring is substantially automatable. |
| Prior risk assessment & new installation advice | 10% | 2 | 0.20 | AUG | Conducts prior risk assessments for new equipment, facility modifications, or new uses of ionising radiation as required under IRR17 Reg 8. Evaluates proposed arrangements against regulatory requirements. Requires understanding of specific equipment, facility constraints, and intended use patterns — context-dependent engineering judgment that AI cannot replicate without full situational awareness. |
| Training, briefings & employer consultation | 5% | 2 | 0.10 | AUG | Delivers radiation protection training to RPSs and classified workers, briefs senior management on regulatory requirements, consults with employers on compliance strategy. Face-to-face advisory interaction — persuading employers to invest in safety measures, explaining regulatory consequences. Not deep relationship work but requires human credibility and interpersonal communication. |
| Contingency planning & emergency advice | 5% | 2 | 0.10 | AUG | Develops contingency plans for radiation accidents, advises on emergency procedures, reviews arrangements for source loss, spillage, and overexposure events. Requires scenario-based judgment about real-world emergency response in specific facilities. AI can assist with scenario modelling but emergency planning requires understanding of local conditions, available resources, and realistic response capabilities. |
| Total | 100% | 2.40 |
Task Resistance Score: 6.00 - 2.40 = 3.60/5.0
Displacement/Augmentation split: 10% displacement, 70% augmentation, 15% not involved, 5% augmentation.
Reinstatement check (Acemoglu): AI creates new tasks — validating AI-generated shielding calculations, auditing automated dose monitoring system outputs, advising on AI-assisted diagnostic equipment radiation characteristics, reviewing digital dosimetry platform configurations, and assessing cybersecurity of networked radiation monitoring systems. These offset efficiency gains in documentation and routine calculations.
Evidence Score
| Dimension | Score (-2 to 2) | Evidence |
|---|---|---|
| Job Posting Trends | 0 | No BLS equivalent (UK statutory role). ~700-1,000 RPA2000-accredited individuals in UK. Indeed UK shows active postings: £48,000+ at AWE, £60,000-£70,000 at Rutherford Appleton Lab, trainee roles at £28,000+. Small occupation with steady but not growing demand. Nuclear new build (Hinkley Point C, Sizewell C) and decommissioning (Sellafield) maintain demand; healthcare expansion (CT/PET growth) adds marginal new positions. |
| Company Actions | 0 | No reports of AI replacing RPAs. Statutory requirement under IRR17 ensures baseline demand. RPA2000 accreditation scheme continues operating normally. Nuclear Industry Association highlights general radiation protection skills gaps. No displacement signals but also no notable expansion signals. |
| Wage Trends | 1 | UK mid-level RPA salaries £48,000-£75,000 depending on sector. Nuclear sector commands highest (£55,000-£85,000+). NHS Band 7/8a equivalent (£45,000-£65,000). Contract/consultancy rates £400-£700+/day. Rutherford Appleton Lab posting at £60,000-£70,000. Wages growing above inflation, reflecting skills shortage in niche statutory profession. |
| AI Tool Maturity | 1 | No AI tools perform statutory RPA advisory functions. Shielding calculation software (e.g., MCNP, BEAMnrc) automates computation but engineer sets up and validates. Dose management platforms (Mirion, Landauer) automate tracking. No production-ready AI tool conducts site inspections, interprets IRR17 for novel situations, or provides statutory radiation protection advice. Industry AI adoption remains minimal in radiation protection advisory work. |
| Expert Consensus | 0 | SRP (Society for Radiological Protection) consensus: augmentation, not displacement. IRR17 statutory framework mandates human RPA appointment. HSE guidance explicitly requires competent human advisers. No expert sources predict AI replacing RPAs. However, limited specific commentary on AI impact — this is a small, specialised profession that receives little AI-impact analysis. |
| Total | 2 |
Barrier Assessment
Reframed question: What prevents AI execution even when programmatically possible?
| Barrier | Score (0-2) | Rationale |
|---|---|---|
| Regulatory/Licensing | 2 | IRR17 Regulation 14 mandates appointment of an RPA who must be accredited by a body recognised by HSE — currently RPA2000 is the sole recognition scheme. RPA2000 accreditation requires demonstrated competence, CPD, and periodic revalidation. No legal pathway for AI to hold RPA2000 accreditation or satisfy the statutory requirement for an appointed RPA. HSE enforcement action for failure to appoint a competent RPA. |
| Physical Presence | 2 | Site inspections of controlled and supervised areas are fundamental to the statutory advisory role. Must physically verify shielding installations, check source stores, inspect workplace arrangements, and assess real-world conditions against documented procedures. Works in hospital departments, nuclear facilities, industrial sites, and research laboratories. Cannot be performed remotely or by AI. |
| Union/Collective Bargaining | 0 | RPAs are typically not unionised. May benefit indirectly from Prospect (union for professionals in nuclear sector) but this does not materially protect the RPA role specifically. |
| Liability/Accountability | 2 | Statutory advisory liability — if RPA advice is negligent and results in worker overexposure, public dose exceedance, or environmental contamination, consequences include HSE prosecution, loss of RPA2000 accreditation, civil liability, and potential criminal charges under IRR17 and the Health and Safety at Work Act 1974. RPAs carry professional indemnity insurance for this reason. AI has no legal personhood to bear statutory advisory liability. |
| Cultural/Ethical | 2 | Radiation protection has deep cultural commitment to human oversight driven by historical incidents (Chernobyl, Goiania, Litvinenko). Workers, employers, and regulators expect qualified human advisers making radiation safety judgments. The ALARP principle requires nuanced human judgment about what is "reasonably practicable." No societal acceptance of AI making statutory radiation protection advisory decisions. Nuclear and healthcare safety cultures strongly resist delegating safety judgments to automated systems. |
| Total | 8/10 |
AI Growth Correlation Check
Confirmed 0 (Neutral). RPA demand is driven by IRR17 statutory requirement — every UK employer using ionising radiation above exemption levels must appoint an RPA. This is regulatory-driven, not AI-driven. AI growth has no material impact on the number of organisations using ionising radiation or the regulatory requirement for RPA appointment. The nuclear renaissance (Sizewell C, SMRs) provides a positive demand driver but is energy-policy-driven, not AI-driven. Not +1 because the relationship between AI growth and RPA demand is negligible.
JobZone Composite Score (AIJRI)
| Input | Value |
|---|---|
| Task Resistance Score | 3.60/5.0 |
| Evidence Modifier | 1.0 + (2 x 0.04) = 1.08 |
| Barrier Modifier | 1.0 + (8 x 0.02) = 1.16 |
| Growth Modifier | 1.0 + (0 x 0.05) = 1.00 |
Raw: 3.60 x 1.08 x 1.16 x 1.00 = 4.5101
JobZone Score: (4.5101 - 0.54) / 7.93 x 100 = 50.1/100
Zone: GREEN (Green >= 48)
Sub-Label Determination
| Metric | Value |
|---|---|
| % of task time scoring 3+ | 45% |
| AI Growth Correlation | 0 |
| Sub-label | Green (Transforming) — AIJRI >= 48 AND >= 20% of task time scores 3+ |
Assessor override: Score adjusted from 50.1 to 55.7. The formula underweights the statutory nature of this role. Unlike most professions where barriers are defensive (preventing AI from taking tasks humans currently do), the RPA barrier is constitutive — the role exists because of regulation. IRR17 does not merely regulate an RPA's work; it creates the role itself. This is stronger than Process Safety Engineer (60.8, barrier 8/10) where the engineer exists independently of OSHA PSM and the regulation adds protection. An RPA exists only because IRR17 requires one. Additionally, the small profession size (~700-1,000) means the evidence score is artificially low — there simply is not enough labour market data to score higher, but the demand is structural and non-discretionary. The override of +5.6 points (to 55.7) places the role between Health Physics Technician (48.5 — lower because hands-on execution rather than advisory authority) and Nuclear Engineer (58.6 — higher because of larger workforce, stronger evidence base, and AI growth correlation). This calibration is honest.
Assessor Commentary
Score vs Reality Check
The 55.7 score sits 7.7 points above the Green boundary. This is not borderline — removing all barriers (modifier drops to 1.00) would yield approximately 43.2 (Yellow), so barriers are load-bearing for this role's Green classification. This is appropriate: the RPA role is entirely barrier-dependent. Without IRR17, there would be no statutory requirement for RPAs. The regulation is not at risk — ionising radiation legislation has been strengthened, not weakened, over decades (from IRR 1985 to IRR 1999 to IRR 2017). The probability of deregulation is effectively zero given the health consequences of uncontrolled radiation exposure.
What the Numbers Don't Capture
- Cross-sector versatility as career insurance. RPAs work across nuclear, healthcare, industrial, and research sectors. A single RPA can hold multiple scopes of accreditation (medical + industrial + nuclear). This provides exceptional career resilience — if one sector contracts, the RPA transfers to another. No single-sector disruption can eliminate demand for the profession.
- Consultant RPA model insulates from employer headcount decisions. Many RPAs work as independent consultants or through specialist consultancies (PHE Associates, RPA Associates, Nuvia, etc.), serving multiple employers simultaneously. This consulting model is resilient to any single employer's cost-cutting because each appointment is a legal compliance requirement, not a discretionary hire.
- Ageing workforce creating succession pressure. The RPA profession skews older. RPA2000 accreditation requires significant post-degree experience, meaning most RPAs are 35+. Retirement-driven attrition is creating supply shortage in a profession where training pipeline takes 5-10 years.
Who Should Worry (and Who Shouldn't)
RPAs with active RPA2000 accreditation conducting site inspections, shielding assessments, dose investigations, and providing face-to-face statutory advice are in one of the most AI-resistant advisory positions in the UK. The statutory appointment requirement, physical inspection mandate, and advisory liability create a triple lock that AI cannot breach under current or foreseeable law. Those most exposed are RPAs who have drifted into purely desk-based compliance documentation roles — writing risk assessments, maintaining records, and producing audit reports without site visits or direct employer consultation. This administrative work is increasingly automatable. The differentiator is whether you are the appointed RPA providing statutory advice (protected) or a radiation protection administrator producing documents (exposed).
What This Means
The role in 2028: RPAs will use AI-enhanced shielding calculation tools, automated dose trending platforms, and AI-generated draft compliance documentation. Digital dosimetry systems will automate routine dose record management. But the core statutory function — conducting site inspections, providing expert radiation protection advice, investigating dose incidents, and bearing personal advisory liability under IRR17 — remains entirely human. The RPA becomes more analytically powerful while remaining personally accountable.
Survival strategy:
- Maintain RPA2000 accreditation and expand scope. Multi-scope accreditation (medical + industrial + nuclear) maximises career flexibility. The accreditation is your statutory moat — without it, you cannot be appointed as an RPA.
- Stay on-site, not behind a desk. Maximise time on site inspections, shielding assessments, and face-to-face employer consultation. The RPA who inspects facilities and advises employers in person is irreplaceable; the one who only writes risk assessments from a desk is doing automatable work.
- Add RWA (Radioactive Waste Adviser) qualification. Dual RPA/RWA accreditation is increasingly valuable as nuclear decommissioning and waste management demand grows. The RWA role has equivalent statutory protection under the Environmental Permitting Regulations.
Timeline: 5-7+ years. IRR17 statutory requirement + RPA2000 accreditation barrier + physical inspection mandate + advisory liability provide strong structural protection. AI transforms calculations and documentation but cannot replace the human appointed as RPA under regulation.
