Role Definition
| Field | Value |
|---|---|
| Job Title | Physicist |
| Seniority Level | Mid-Level |
| Primary Function | Conducts original research across physics subfields (condensed matter, particle, quantum, optics, astrophysics, plasma, atomic/molecular). Designs and runs experiments using specialised apparatus (accelerators, cryostats, lasers, detectors), develops theoretical models, analyses complex datasets, publishes peer-reviewed papers, and competes for grant funding. Typically holds a postdoctoral or early staff scientist position at a university, national laboratory, or industry R&D facility. |
| What This Role Is NOT | NOT an astronomer (observatory-based, separate SOC 19-2011, scored 45.2 Yellow). NOT a physics teacher (postsecondary, separate SOC). NOT a junior postdoc executing protocols designed by a PI. NOT a senior principal investigator directing a large research programme. NOT an applied engineer using physics principles in product development. |
| Typical Experience | PhD in physics + 2-8 years postdoctoral or staff experience. No formal licensure — credentialling is via publication record, grant success, and facility access allocations. |
Seniority note: Junior postdocs focused primarily on data processing and protocol execution would score lower Yellow. Senior PIs and lab directors who set research agendas, lead international collaborations, and bear institutional accountability would score higher Green (~58-62).
Protective Principles + AI Growth Correlation
| Principle | Score (0-3) | Rationale |
|---|---|---|
| Embodied Physicality | 1 | Experimental physicists operate specialised apparatus — particle accelerators, vacuum chambers, cryogenic systems, optical tables, cleanrooms. But work occurs in structured laboratory environments, not unstructured physical settings. Computational and theoretical physicists are fully desk-based. |
| Deep Interpersonal Connection | 0 | Collaborative research matters, but the core value is the science, not human-to-human relating. Transactional professional collaboration, not trust-based connection. |
| Goal-Setting & Moral Judgment | 2 | Significant judgment in choosing which physical questions to investigate, designing novel experiments, interpreting ambiguous or unexpected results, and deciding when a finding is publishable. Defines research direction within their subfield. Operates at the frontier where no playbook exists. |
| Protective Total | 3/9 | |
| AI Growth Correlation | 0 | AI adoption neither creates nor destroys physicist demand. AI accelerates simulation and data analysis but does not change whether humans are needed to do the physics. Demand driven by government R&D funding (DOE, NSF, DARPA), industry investment, and fundamental scientific questions. Net effect: neutral. |
Quick screen result: Protective 3 + Correlation 0 = Likely Yellow/Green boundary. Proceed to quantify.
Task Decomposition (Agentic AI Scoring)
| Task | Time % | Score (1-5) | Weighted | Aug/Disp | Rationale |
|---|---|---|---|---|---|
| Experimental design & apparatus development | 20% | 2 | 0.40 | AUG | Designing novel experiments — building detector arrays, configuring laser systems, fabricating quantum devices — requires deep physical intuition and engineering judgment. AI assists with parametric optimisation but the physicist defines what to measure and how. |
| Laboratory experimentation & data collection | 15% | 2 | 0.30 | AUG | Operating accelerators, cryostats, optical benches, and cleanroom fabrication equipment. Physical dexterity in unstructured lab settings. AI-guided instruments (e.g., Berkeley Lab beamline tuning) assist but the physicist troubleshoots, adapts, and operates in real time. |
| Data analysis & computational modelling | 20% | 3 | 0.60 | AUG | AI handles significant sub-workflows — ML-based particle classification, automated spectral analysis, simulation acceleration (NVIDIA Apollo 10x speedup). But the physicist leads analysis design, validates physical significance, interprets edge cases, and determines what the data means. Human-led, AI-accelerated. |
| Theoretical development & hypothesis generation | 20% | 1 | 0.20 | NOT | The irreducible core — developing new theoretical frameworks, deriving novel physics from first principles, connecting experimental anomalies to fundamental theory. Genuine novelty creation. AI has no capacity to decide which physical questions matter or to generate original theoretical insights. |
| Scientific writing & peer review | 10% | 3 | 0.30 | AUG | AI drafts sections, generates figures, assists with literature synthesis. But the scientific narrative, interpretation of results, and peer review judgment remain human. 53% of physics peer reviewers now use AI tools, but the intellectual contribution must be human. |
| Grant writing & funding acquisition | 10% | 2 | 0.20 | AUG | AI assists with literature review and section drafting. But identifying knowledge gaps, articulating scientific significance, and persuading expert panels requires deep domain judgment. DOE/NSF review panels evaluate investigator insight and novelty. |
| Teaching, mentoring & collaboration | 5% | 1 | 0.05 | NOT | Mentoring graduate students, teaching courses, leading research group meetings, building international collaboration networks. Human relationships and pedagogical judgment. |
| Total | 100% | 2.05 |
Task Resistance Score: 6.00 - 2.05 = 3.95/5.0
Displacement/Augmentation split: 0% displacement, 75% augmentation, 25% not involved.
Reinstatement check (Acemoglu): AI creates substantial new tasks: validating AI-accelerated simulation results against experimental data, designing ML training sets for detector systems, interpreting AI-discovered anomalies in particle collision data, developing "physical AI" workflows that integrate ML with physics-informed models, and operating DOE Genesis Mission AI tools for automated discovery. The role is expanding, not shrinking.
Evidence Score
| Dimension | Score (-2 to 2) | Evidence |
|---|---|---|
| Job Posting Trends | 0 | BLS projects 4% growth for physicists and astronomers 2024-2034 (about average). 24,600 physicists employed, ~1,800 annual openings. Market is stable, not surging or declining. Quantum computing and national lab expansion create pockets of demand but overall posting volumes are flat. |
| Company Actions | 0 | No reports of national labs or research institutions cutting physicist positions citing AI. DOE Genesis Mission (late 2025) embeds frontier AI tools in national labs — augmenting, not replacing. Quantum computing startups hiring physicists. No net headcount change signal. |
| Wage Trends | 1 | BLS median $166,290 (May 2024) — among the highest in physical sciences. Quantum computing and AI-physics hybrid roles command premiums. Wages growing modestly above inflation, with industry (tech, defence, national labs) outpacing academia. |
| AI Tool Maturity | 1 | Powerful tools augment but do not replace: NVIDIA Apollo (10x simulation speedup), ML-stabilised laser accelerators (Berkeley BELLA), AI-guided beamline optimisation (ALS-U), AlphaProof/Lean for proof formalisation. Tools handle computational volume no human could process manually. No tool replaces the core research function. |
| Expert Consensus | 1 | Broad agreement that AI transforms physics research methods but does not displace physicists. AI Scientist-v2 autonomously generates papers but requires human validation and experimental grounding. Physicists who integrate AI are more productive; those who do not face competitive disadvantage. Consensus: augmentation, not displacement. |
| Total | 3 |
Barrier Assessment
Reframed question: What prevents AI execution even when programmatically possible?
| Barrier | Score (0-2) | Rationale |
|---|---|---|
| Regulatory/Licensing | 1 | No formal licensure, but facility access (particle accelerators, nuclear reactors, classified defence labs) requires qualified human researchers. DOE/NSF mandate human PIs on grants. NRC oversight for nuclear/radiation research. Security clearances required for national lab and defence work. |
| Physical Presence | 1 | Experimental physicists require physical presence — operating particle accelerators, aligning optical systems, fabricating quantum devices in cleanrooms, maintaining cryogenic equipment. Structured lab environments. Computational/theoretical work is fully remote-capable. |
| Union/Collective Bargaining | 0 | Scientists are not meaningfully unionised. Some postdoc unions at specific universities. No structural protection against AI-driven role changes. |
| Liability/Accountability | 0 | Low personal liability — incorrect physics findings do not endanger lives or create legal consequences. Reputational risk (retraction, loss of grant funding) exists but is not a structural barrier in the AIJRI sense. |
| Cultural/Ethical | 1 | Scientific community values human-driven discovery. Publications, tenure, prizes (Nobel, APS fellowships) built around human intellectual contribution. AI-generated papers without meaningful human contribution face rejection. Physics culture requires human accountability for research claims. |
| Total | 3/10 |
AI Growth Correlation Check
Confirmed at 0 (Neutral). AI adoption in physics is substantial — ML for detector data, simulation acceleration, autonomous experimental tuning — but this creates efficiency gains within existing teams, not demand for more physicists. The field's size is constrained by government R&D budgets (DOE, NSF, DARPA), facility access, and faculty/staff lines, not by computational bottlenecks. Quantum computing creates some physicist demand (quantum hardware, error correction), but this is a niche within the broader occupation. Not Accelerated Green (no recursive AI dependency). Not negative (AI makes physicists more productive, not obsolete).
JobZone Composite Score (AIJRI)
| Input | Value |
|---|---|
| Task Resistance Score | 3.95/5.0 |
| Evidence Modifier | 1.0 + (3 × 0.04) = 1.12 |
| Barrier Modifier | 1.0 + (3 × 0.02) = 1.06 |
| Growth Modifier | 1.0 + (0 × 0.05) = 1.00 |
Raw: 3.95 × 1.12 × 1.06 × 1.00 = 4.6894
JobZone Score: (4.6894 - 0.54) / 7.93 × 100 = 52.3/100
Zone: GREEN (Green >=48, Yellow 25-47, Red <25)
Sub-Label Determination
| Metric | Value |
|---|---|
| % of task time scoring 3+ | 30% |
| AI Growth Correlation | 0 |
| Sub-label | Green (Transforming) — AIJRI >=48 AND >=20% task time scores 3+ |
Assessor override: None — formula score accepted. The 52.3 score calibrates well against comparators: higher than Astronomer (45.2 Yellow) due to stronger task resistance (3.95 vs 3.60) — physicists have deeper experimental design and apparatus work that astronomers increasingly outsource to automated pipelines. Lower than Medical Scientist (54.5) due to weaker evidence (+3 vs +5) and barriers (3/10 vs 4/10) — medical scientists benefit from stronger BLS growth (9% vs 4%) and FDA/IRB regulatory mandates that physicists lack. The label is honest.
Assessor Commentary
Score vs Reality Check
The 52.3 sits 4.3 points above the Green boundary (48) — inside Green but not deeply so. Stripping barriers entirely (modifier drops to 1.00) yields 49.4 — still Green by 1.4 points, meaning the classification is not barrier-dependent. The strong task resistance (3.95) reflects the genuinely irreducible nature of hypothesis generation, theoretical development, and experimental design. What keeps the score modest is the combination of weak barriers (3/10 — no licensing, no liability, modest cultural protection) and moderate evidence (+3 — stable but not growing market). The 4.3-point margin above the boundary is documented but does not warrant an override.
What the Numbers Don't Capture
- Subfield divergence. Condensed matter experimentalists designing and fabricating novel quantum devices in cleanrooms live in a different zone than computational physicists whose work overlaps heavily with ML capabilities. The 3.95 average masks a genuine bimodal split between deeply physical experimental work (score 1-2) and computational modelling (score 3).
- Funding dependency. Physicist employment tracks government R&D budgets more than AI capability curves. A Congressional increase in DOE/NSF funding would move the evidence score regardless of AI. Federal budget sequestration or political shifts in science funding priorities would do the opposite. The current 4% growth projection assumes stable funding — a fragile assumption.
- Quantum computing as demand catalyst. Quantum hardware, error correction, and quantum information science create growing demand for physicists specifically — not computer scientists. This niche is not large enough to shift the overall occupation evidence score but represents a genuine growth vector for physicists who specialise.
- PhD as implicit barrier. The PhD requirement (5-7 years) functions as a de facto entry barrier not captured in the formal barrier score. AI cannot earn a PhD, and the scientific community uses the credential as a proxy for demonstrated research capability.
Who Should Worry (and Who Shouldn't)
If you are an experimental physicist who designs novel apparatus, runs physical experiments, and develops new theoretical frameworks, you are doing work AI cannot replicate. The "Transforming" label means your data analysis pipeline, simulation workflow, and literature review process are changing — but the core intellectual and experimental work is protected. Most protected: condensed matter experimentalists fabricating novel materials and devices, particle physicists designing detector systems, and quantum experimentalists building hardware. More exposed: computational physicists whose primary output is simulation code and numerical modelling — this work overlaps significantly with AI capabilities. The single biggest separator is whether you are generating new physics or processing existing data. The theorist and experimentalist are protected; the human computing engine is increasingly augmented to the point where fewer are needed per unit of output.
What This Means
The role in 2028: The surviving mid-level physicist uses AI as standard research infrastructure — ML-accelerated simulations, automated detector calibration, AI-guided experimental optimisation, and literature synthesis tools. One researcher with AI tools produces what two or three produced manually in 2020. But the physicist still generates every hypothesis, designs every experiment, builds every novel apparatus, validates every AI prediction against physical reality, and bears accountability for every published result.
Survival strategy:
- Build computational-experimental integration skills. The physicist who bridges hands-on lab work with ML-powered analysis is most valuable. Python, PyTorch, and physics-informed neural networks are now baseline competencies.
- Specialise where physicality and novelty intersect. Quantum device fabrication, novel detector design, ultrafast laser experiments — areas where physical intuition and manual dexterity combine with theoretical depth have the strongest moats.
- Develop AI-physics hybrid expertise. DOE Genesis Mission tools, AI-guided beamline operations, and ML-accelerated simulation workflows represent where the field is heading. Physicists who design AI training sets and validate AI-generated predictions position for the transformed role.
Timeline: 10-15+ years. Protected by the irreducibility of the scientific method (hypothesis, experiment, interpretation), PhD training pipeline (10+ years), funding-constrained market size, and the expanding frontier of unanswered physics questions. Data analysis and simulation workflows transform within 3-5 years; the core research function persists.
