AI as Research Notebook, Tutor, Coding Partner, and Critic

A practical operating model for using AI day-to-day while studying and practicing materials science.

Four roles

AI becomes more useful when its role is explicit. For a materials scientist, four roles cover most responsible use: notebook, tutor, coding partner, and critic.

1. Research notebook

Use AI to turn messy reading into structured notes. Ask for a table with the paper's objective, material system, processing route, characterization methods, key findings, limitations, and follow-up questions.

Then verify every important claim against the source. The goal is not to let AI read for you. The goal is to make your own reading easier to compare later.

2. Tutor

Use AI to explain concepts at different levels. Ask for an undergraduate explanation, then a graduate-level explanation, then a mechanism-focused explanation with equations. This is useful for topics like diffusion, creep, corrosion, dislocations, fracture toughness, crystallinity, phase diagrams, and irradiation damage.

Good prompt: Explain this concept, then ask me five questions that would reveal whether I understand it.

3. Coding partner

Use AI to write first drafts of Python scripts for plotting, fitting, cleaning data, and building notebooks. Always run the code on simple known inputs. Check axes, units, labels, and edge cases.

Good prompt: Write a Python script to plot stress-strain data from a CSV, calculate modulus over a user-selected strain range, and label all units clearly. Include a small synthetic example so I can test it.

4. Critic

Before sharing work, ask AI to review your argument. Tell it to be strict. Ask what a senior engineer, professor, or reviewer would challenge. Ask which claim is weakest and what evidence would strengthen it.

This role may be the most valuable because it changes AI from a generator into a pressure test.

The rule

Use AI to improve the quality of your thinking. Do not use it to bypass thinking. The difference shows.

Sources and starting points

• NIST Data and AI-Driven Materials Science Group: https://www.nist.gov/mml/mmsd/data-and-ai-driven-materials-science-group
• Materials Project documentation: https://docs.materialsproject.org/
• pymatgen documentation: https://pymatgen.org/
• matminer documentation: https://hackingmaterials.lbl.gov/matminer