If you've started using artificial intelligence and have ever thought “"Okay... this isn't what I wanted"”, Relax: it's not that the AI fails, it's that You don't speak the same language yet. And this is where a key word comes into play: prompt.
Because yes, using AI isn't just about copying and pasting random questions. It's about knowing how to ask. Just like in science.
What is a prompt?
A prompt It's basically the instruction you give to an artificial intelligence to generate a response. It's what you write before the AI starts working: a question, a command, a context, a goal… or all of the above.
Simply put: The prompt is how you communicate with the AI.
And here's the important part: The quality of the response depends directly on the quality of the prompt.
- If you ask a vague question, you'll get a vague answer.
- If you provide context, clear objectives, and limits, AI responds much better.
It's like asking a lab partner for help:
– “Hey, help me with this”
Vs
– “Hey, can you review this section of the method and tell me if it’s clear and reproducible?”
Do you see the difference? Well, the same thing happens with AI.
How a prompt works
AI doesn't "understand" like a person, but detects patterns in language and it responds based on what you ask it to. That's why when you write a prompt, you are defining:
- What do you want me to do? (summarize, explain, write, compare…)
- What topic?
- At what level of depth
- In what format
- For whom (students, experts, outreach, etc.)
A good prompt usually includes several of these elements. For example:
“Summarize this scientific article, highlighting the hypothesis, methodology, and main results, using a clear and concise tone, intended for researchers.”
The result? Much more useful than a simple "summary".
Furthermore, prompts are not something rigid. They are adjusted, improved, and refined.. Just like an experiment: you test, observe the results, and correct.
Examples of basic prompts and their results
Let's get down to brass tacks, which is where this makes sense.
Basic Prompt
“Explain to me what photosynthesis is.”
Result:
A general explanation, correct, but quite basic.
Improved Prompt
“"Explain photosynthesis to me clearly, using simple language, but without losing scientific rigor."”
Result:
A more careful, precise and adapted explanation.
Prompt for research
“Summarize this paper, highlighting the study's objective, methodology, main results, and limitations.”
Result:
A structured summary that saves you time without losing key information.
Prompt for scientific writing
“Rewrite this paragraph from the introduction to make it clearer and more fluid, while maintaining an academic tone.”
Result:
A cleaner, more coherent text, ready for final review.
Prompt with role
“Act as a reviewer for a scientific journal and point out possible methodological weaknesses in this section.”
Result:
Critical feedback that helps you improve your manuscript before submitting it.
Using effective prompts is a skill. It's not magic, intuition, or "having a knack for AI." It's practice and method.
If you want to delve deeper into how to structure prompts, adapt them to scientific tasks, and truly get the most out of them (beyond the typical "give me a summary"), there are specific resources such as the course The art of asking, where it is taught prompt engineering applied to real-world contexts, including research.
Not to depend on AI, but to Use AI judiciously.
Biotechnologist and biostatistician specializing in data analysis and artificial intelligence applied to research. Founder of Maxymia, combines science and technology to develop solutions and training in automation for the scientific field.
