AI helps with lab reports, statistics and reading papers. It has no place in the lab itself, though: take procedures and safety from the instructions and the safety data sheet, and never change measured data based on AI. Prompts for the lab, statistics, papers and exams. As of 26 October 2026.
How universities handle AI and what doesn't belong in it in general is on the university page. Here's what applies on top of that in the natural sciences.
AI reads your report and calculations like a strict assistant. The data and the text stay yours.
<report> [paste your report] </report> Check my lab report in [analytical chemistry]: - Does it have all the parts (aim, principle, method, results, discussion, conclusion)? - Are the units, significant figures and rounding right? - Does the conclusion follow from the measured data, or does it claim more than the data show? - What's missing from the discussion (sources of error, comparison with the reference value)? Don't change or calculate any measured data. Don't rewrite the text, just write your comments.
PD2V Web <calculation>
[paste your calculation, such as diluting a solution or the yield of a reaction]
</calculation>
Check my calculation step by step. At each step, check the units (dimensional analysis) and whether the number is in a sensible order of magnitude.
Find the first mistake and explain it. Don't calculate the rest for me.
AI suggests the test and explains the result. The software does the calculating, not AI.
My experiment: [what you measure, how many groups, how many measurements per group, whether the measurements are paired]. The data: [type of data and distribution, or paste a sample]. 1. Suggest a suitable statistical test and explain why. 2. Say what assumptions the test has and how I check them. 3. Suggest an alternative test in case the assumptions don't hold. 4. Write code in [R / Python with SciPy] that runs the test. Don't make up results, I'll run the test myself.
<output>
[paste the output from R, Python, SPSS or Excel]
</output>
Explain this output to me: what each value is, what the p-value and confidence interval mean in my case, and how to write the result in my report in one sentence.
Also say what I can't claim from the result.
Work only from the numbers in the output.
A graph from a paper and designing your own experiment. How to read a whole paper in a foreign language is on the university page.
Upload the image of the graph along with its caption.
I'm uploading a graph from a scientific paper. Before you explain it, ask me: 1. What's on the axes and in what units? 2. What do the error bars mean (standard deviation, standard error of the mean, confidence interval)? 3. What's the main trend and is the difference between the groups large? Then correct me and fill in what I missed. If something in the image can't be read precisely, say so and don't guess.
I'm designing an experiment: [hypothesis, what I'll change, what I'll measure, how many replicates].
You're the opponent at a seminar. Find the weak points: missing controls, variables I don't control, too few replicates, measurement errors. For each comment, suggest how to fix it in the design.
Don't write a whole new design, I want to adjust it myself.
Processes step by step, identification quizzes and getting ready for fieldwork.
Quiz me on [the SN2 reaction mechanism / the Krebs cycle / photosynthesis]. I want to be able to explain it step by step.
Give me just the start and let me describe the next step. When I get it wrong, give me a hint, not the whole answer. At the end, show me the whole process and mark the steps where I made mistakes.
If you're not sure about a step, say so.
Don't rely on AI to identify from a photo; it can mix up similar species. Take photos from an atlas or a key.
<notes> [paste a list of species or minerals and their features from your notes] </notes> Quiz me on identifying [trees / birds / minerals] by the features in my notes. Describe the features without the name and I'll guess. Ask one at a time. Take the features only from my notes. At the end, tell me which ones I mix up.
I'm going on a field trip in [botany / geology / ecology] to [a floodplain forest / a quarry / a mountain meadow]. Tell me: what to review, what to bring, how to keep a field notebook and what to watch out for for safety. Keep it general. My teacher will give me the specific sites and instructions.