Jobs › Life, Physical, and Social Science
AI exposure report
Will AI replace Physicists?
Many of your tasks are within reach of AI. The role will change, so it pays to act now.
Task exposure to today's AI
High exposure
This is the occupation's typical task mix. Your own score can be up to 30 points higher or lower depending on how you work: routine share, people contact, judgment and how much you already use AI.
Get your personal scoreWill AI replace Physicists?
Not entirely. Today's AI tools can take on about 50% of the typical work in this occupation, mostly: perform complex calculations as part of the analysis and evaluation of data, using computers. The parts that stay human include: Observe the structure and properties of matter, and the transformation and propagation of energy, using equipment such as masers, lasers, and telescopes, to explore and identify the basic principles governing these phenomena.; Analyze data from research conducted to detect and measure physical phenomena.; Develop theories and laws on the basis of observation and experiments, and apply these theories and laws to problems in areas such as nuclear energy, optics, and aerospace technology..
People who learn to use AI for the automatable parts keep the edge. The 3-phase plan below shows where to start.
Which tasks AI can take on
The largest tasks in this occupation, their share of the working week, and how much of each today's AI tools can do.
+ 4 more tasks (29% of the week) in the full test.
Your human edge
- Observe the structure and properties of matter, and the transformation and propagation of energy, using equipment such as masers, lasers, and telescopes, to explore and identify the basic principles governing these phenomena.
- Analyze data from research conducted to detect and measure physical phenomena.
- Develop theories and laws on the basis of observation and experiments, and apply these theories and laws to problems in areas such as nuclear energy, optics, and aerospace technology.
AI tools to try
ChatGPT / Claude for literature and writingAI research tools (e.g. Elicit, Consensus, Semantic Scholar)Copilot for data analysis code (Python, R)Related roles
- Mathematicians
- Astronomers
- Data Scientists
A 3-phase AI learning plan
Phase 1
Faster literature and writing
- Literature search and summaries with AI, always checked against the papers.
- Drafts of methods, reports and grant text.
- Reference management and formatting.
Project
Use AI to summarise 20 papers for your next manuscript's introduction, verify every claim and citation against the PDFs, and log how many hours the draft saved.Phase 2
AI for data analysis
- Analysis code in Python or R written with AI help.
- Visualisation and statistics explained step by step.
- Reproducibility: version code and prompts.
Project
Rebuild one past analysis pipeline in Python with AI help, add fits, error bars and plots, and version the code plus prompts in a repo that reproduces your published figures.Phase 3
AI as a research method
- Classification, extraction or modelling with AI on your data.
- Validate against human labels or ground truth.
- Ethics, data protection and disclosure of AI use.
Project
Train an AI classifier to flag events or spectral features in 1,000 of your measurements, compare its labels with your manual ground truth, and report accuracy plus an AI-use disclosure.