Occupation Report · Engineering
Physicists investigate the fundamental laws governing matter, energy, space, and time — applying mathematical and experimental frameworks across fields from condensed matter and quantum mechanics to astrophysics, particle physics, and applied photonics. The role combines theoretical modelling, complex data analysis, experimental design, and scientific communication. AI is accelerating literature synthesis and certain data processing tasks, while the deep theoretical reasoning, novel experimental design, and physical intuition that define frontier physics research remain robustly human.
Last updated: Mar 2026 · Based on O*NET, Frey-Osborne, and live labour market data
AI Exposure Score
Window to Act
AI is accelerating peripheral physics tasks such as literature synthesis and data pipeline automation but makes limited inroads into the theoretical creativity and physical experimental design at the heart of the role. Structural displacement is unlikely before the mid-to-late 2030s, with meaningful change concentrated in the computational and administrative periphery.
vs All Workers
Physicists sit in the bottom third of AI displacement risk across the UK workforce, protected by the deep theoretical reasoning, novel experimental design, and mathematical creativity that characterise the profession and that AI systems cannot yet replicate reliably.
Physics sits well below average on AI displacement risk. Literature synthesis and data pipeline automation are increasingly AI-assisted, but the theoretical creativity, novel experimental design, and physical laboratory work that define the profession are strongly human-dependent.
| Task | Risk Level | AI Tools Doing This | Exposure |
|---|---|---|---|
|
Literature Review & Theoretical Research
Systematically reviewing physics literature across arXiv, journal databases, and conference proceedings to understand the current theoretical framework and identify new research directions.
|
High | Elicit, Semantic Scholar, arXiv AI tools, SciSpace, Perplexity AI |
|
|
Data Analysis & Signal Processing
Analysing experimental datasets from detectors, telescopes, or laboratory instruments — applying signal processing, statistical methods, noise reduction, and pattern recognition to extract physical signals.
|
Medium | ChatGPT Code Interpreter, CERN ROOT with ML extensions, PyTorch, SciPy, MATLAB AI Toolbox |
|
|
Theoretical Modelling & Simulation
Developing mathematical models of physical systems, writing simulation code, and validating theoretical predictions against experimental observations to advance understanding of physical phenomena.
|
Medium | GitHub Copilot, ChatGPT (equation derivation support), Wolfram Mathematica AI, MATLAB Copilot |
|
|
Scientific Writing & Peer Review
Writing peer-reviewed manuscripts, preparing grant applications to STFC and EPSRC, contributing to review articles, and critically evaluating manuscripts submitted to physics journals.
|
Medium | Writefull, PaperPal, ChatGPT, Claude, Grammarly Business |
|
|
Experimental Design & Apparatus Development
Designing experiments to test specific physical hypotheses, specifying apparatus requirements, determining measurement precision, and developing novel instrument configurations.
|
Low | ChatGPT (design review support), COMSOL Multiphysics with AI features |
|
|
Physical Experimentation & Instrument Operation
Building experimental apparatus, calibrating instruments, performing measurements in controlled environments, and collecting primary data in laboratory or large-facility settings such as synchrotrons and particle accelerators.
|
Low | Automated data acquisition systems (NI LabVIEW), robotic sample handlers, physics instrument AI diagnostics |
|
|
Teaching, Supervision & Expert Consultation
Delivering undergraduate and postgraduate teaching, supervising doctoral students, and providing expert advisory services to industry, government, or legal clients on technically complex physical questions.
|
Low | ChatGPT (teaching material drafting), Notion AI (lecture notes), AI tutoring platforms as supplementary tools |
AI has entered physics research primarily through literature synthesis and data pipeline tools, with the theoretical creativity and experimental ingenuity that define the discipline proving far more resistant to automation than in most scientific fields.
2018–2023
Machine learning enters astrophysics and particle physics data
The 2018–2023 period saw machine learning achieve notable results in physics data contexts: neural networks classifying galaxy morphologies, ML-based event selection in particle physics experiments at CERN, and deep learning applied to gravitational wave detection. Literature tools like Semantic Scholar and Elicit improved search quality across arXiv. Core theoretical physics work and experimental apparatus design remained untouched by AI, and the global physics community continued to grow.
2024–2026
LLMs accelerate literature synthesis and code writing
By 2025, most physically active researchers use LLMs for literature synthesis, code generation in Python and MATLAB, and first-draft scientific writing. AI-assisted symbolic mathematics tools are improving at equation manipulation. However, generating novel physical theories, designing experiments with genuinely new capabilities, and the physical intuition required to diagnose why an experiment is behaving unexpectedly remain very strongly human capabilities that AI cannot replicate reliably.
2027–2035
Autonomous analysis for routine physics problems
AI will become more capable of handling data analysis and simulation tasks for well-defined physics problems, allowing physicists to focus on frontier questions. Fields with large, structured datasets — particle physics, observational astronomy — will see the most AI integration. Theoretical physics, novel experimental design, and the physical creativity required to conceive new paradigms will remain highly resilient to automation, sustaining strong demand for skilled physicists across academia, defence, finance, and advanced manufacturing.
Physicists sit well below average on AI displacement risk, protected by the depth of theoretical reasoning and experimental creativity the role demands. The contrast with more automatable data analysis and administrative roles is particularly stark.
More Exposed
Data Analyst
62/100
Data analysts processing structured business datasets face a significantly higher proportion of automatable tasks than physicists whose work requires deep theoretical foundations and novel experimental design.
This Role
Physicist
38/100
Literature synthesis and data pipeline automation are AI-assisted, but theoretical modelling, experimental design, and physical laboratory work require the depth of scientific expertise that AI cannot yet replicate.
Similar Risk
Research Scientist
34/100
Broader research scientists share physics' protection from automation through hypothesis generation and physical experimentation, sitting in the same well-protected bracket.
Much Lower Risk
Doctor
30/100
Clinical medicine's physical examination, patient trust, and life-critical decision-making place it firmly in the well-protected tier, slightly below even physics on the displacement spectrum.
Physicists are exceptionally well-equipped for quantitative careers outside academia, with mathematical modelling, data analysis, and scientific rigour translating powerfully into data science, quantitative finance, and engineering roles.
Path 01 · Adjacent
Aerospace Engineer
↑ 81% skill match
Positive direction
Target role is somewhat more resilient than the source.
You already have: Engineering and Technology, Mathematics, Critical Thinking, Design
You need: Mechanical, Production and Processing, Operations Monitoring, Transportation
Path 02 · Adjacent
Chemical Engineer
↑ 71% skill match
Positive direction
Target role is somewhat more resilient than the source.
You already have: Engineering and Technology, Chemistry, Mathematics, Science
You need: Production and Processing, Mechanical, Operations Monitoring, Troubleshooting
Path 03 · Cross-Domain
Data Science Consultant
↑ 45% skill match
Positive direction
Applies analytical rigor to business problems across industries beyond engineering.
You already have: mathematical modeling, statistical analysis, research methodology, problem-solving, computational skills
You need: machine learning algorithms, business acumen, data visualization tools, industry-specific applications, consulting frameworks
Your personalised plan
Take the free assessment, then get your Physicist Career Pivot Blueprint — a 15-page roadmap with skill gaps, 90-day action plan, salary data, and named employers.
Free assessment · Blueprint: £49 · Delivered within 1–2 business days
Will AI replace Physicists?
AI will not replace physicists in any foreseeable timeframe. The discipline sits among the most AI-resilient in the economy, protected by the depth of mathematical creativity, novel theoretical reasoning, and physical experimental judgment it demands. Peripheral tasks like literature synthesis and code generation are genuinely accelerated by AI, making physicists more productive rather than obsolete. The skills that define excellent physicists — generating new theoretical frameworks, designing experiments that can distinguish competing hypotheses, and interpreting anomalous results — are exactly the capabilities AI systems consistently struggle with most profoundly.
Which Physicist tasks are most at risk from AI?
Literature review and evidence synthesis face the most near-term disruption, with tools like Elicit and Semantic Scholar compressing arXiv searches dramatically. Standard data analysis pipelines using established statistical methods are increasingly AI-assisted through code generation tools like GitHub Copilot. Theoretical modelling for well-understood physical systems can be accelerated by AI-assisted mathematics tools, though genuine novel theory derivation remains a distinctly human activity.
How quickly is AI changing Physicist jobs?
Change is proceeding quickly at the periphery — most actively publishing physicists now use AI tools for literature synthesis and coding — but extremely slowly at the theoretical and experimental core. Even in data-intensive subfields like particle physics and astrophysics, where ML has been adopted enthusiastically, human physicists remain essential for designing experiments, interpreting unexpected results, and constructing the theoretical frameworks that give data meaning. A step-change disruption to core physics work is not expected before the late 2030s at the earliest.
What should Physicists do to stay relevant as AI advances?
Develop strong Python and scientific computing skills to stay productive with AI-augmented analysis platforms. Build familiarity with the AI tools most relevant to your subfield — whether that is ML-based event classification in particle physics, morphology recognition in astrophysics, or AI-assisted materials simulation. Value the uniquely human capabilities that secure physicists' career resilience: theoretical originality, experimental creativity, and the communication of complex physical concepts to non-specialist audiences in industry and policy roles.