Skip to content

A AI · STUDY SKILLS

Traps Workflows Quality check Further reading

EN / DA

A PRACTICAL GUIDE FOR STUDENTS

Learn more
with AI.

Use AI to ask better questions, test your understanding and challenge your own work. This guide gives you practical workflows and prompts you can use right away.

RULE OF THUMB

Use AI to increase the amount of thinking you do yourself.

  • Try it yourself
  • Ask for a hint
  • Explain it in your own words

WHAT THIS GUIDE HELPS YOU DO

A polished output can feel like learning. Competence grows through the work you perform yourself.

When you read, think, solve problems and write a first draft, you build the structures that help new knowledge stick. AI can support this process through questions, feedback, critique and planning.

01

Keep ownership

You make the academic choices and formulate your first attempt.

02

Create productive friction

You use questions and hints to work your way towards understanding.

03

Check confident claims

You verify claims, quotations and references against the syllabus and primary sources.

FOUR COMMON TRAPS

Notice when AI takes over the work that was meant to build your competence.

The traps become easier to spot when you look at your own activity: Who frames the problem, follows the argument and makes the important choices? Each card shows what the trap looks like, why it can be a problem and what you can do instead.

01 · ●

The answer machine

What it looks like

You enter a question, read the answer and move on.

Why this can be a problem

If AI does the work before you have tried yourself, you get less practice thinking your way towards the answer.

Your next move

Answer first. Then use AI to ask a follow-up question or test your explanation.

02 · ●

Summary without thinking

What it looks like

The AI summary replaces your reading of the text.

Why this can be a problem

If AI creates the whole overview for you, you can easily miss the key arguments, connections and questions in the text.

Your next move

Map the argument, hesitation and gaps yourself. Then compare your reading with the AI response.

03 · ●

Delegation as sparring

What it looks like

AI chooses the problem, theory, method and first formulation.

Why this can be a problem

If AI makes the important academic choices for you, it becomes harder to know whether you can explain and defend those choices yourself.

Your next move

State your choices first. Ask AI to critique them, and decide for yourself what you will change.

04 · ●

AI as an authority

What it looks like

A fluent answer is allowed to stand as evidence.

Why this can be a problem

An answer can sound convincing and still be wrong, imprecise or missing important context.

Your next move

Find the original source. Check the quotation, date, concept and context before using it.

01 YOUR FIRST WORKFLOW

THE 10-MINUTE RULE

Give your brain something to work with first.

When you meet a new problem or concept, spend the first ten minutes on your own attempt. AI can then offer a small hint that you work forward from.

00–10 min

Write an attempt

State what you know, where you are stuck and what you want to try.

Next step

Ask for one hint

Give AI your attempt and request the smallest amount of help that moves you forward.

At the end

Explain the solution

Close the chat and explain the method in your own words.

Can you explain the same thing tomorrow without opening the chat?

COPYABLE PROMPTS

Five ways to use AI as a learning partner.

Insert your topic, text or syllabus where you see square brackets. Preserve your own first attempt and always evaluate the response academically.

01 · TUTOR WITH HINTS

Work your way towards the answer

Use this when you are stuck after making your own first attempt.

PROMPT

Act as my tutor in [topic]. I have already tried this: [insert]. Do not give me the full solution. Ask one question at a time, give one small hint, and wait for my response. End by asking me to explain the solution in my own words.

02 · SELF-TEST

Retrieve knowledge from memory

Use this after reading and again a few days later.

PROMPT

Create 8 questions of increasing difficulty about [topic]. Ask one question at a time. Wait for my answer before continuing. Give brief feedback and one improvement hint. At the end, ask me to explain the two areas I found most difficult.

03 · CRITICAL OPPONENT

Test the strength of your argument

Use this after writing your own claim or analysis.

PROMPT

Here is my argument: [insert]. Give me 3 strong counterarguments and ask 2 questions that test my assumptions. Identify the evidence that could change the conclusion. Do not rewrite my text.

04 · REVERSE OUTLINE

See the structure of your draft

Use this when a draft feels unclear or repetitive.

PROMPT

Create a reverse outline of my text: (1) one claim per paragraph, (2) evidence or example, (3) the paragraph’s function, and (4) what is missing. Write no new prose. Give only a diagnosis that I can work from. Text: [insert].

05 · METACOGNITIVE PLAN

Plan around your actual gaps

Use this for revision, exam preparation and reflection.

PROMPT

I am working with [topic/syllabus] and have [time period] available. First ask what I already know and where I feel uncertain. Then identify 3 likely misconceptions and create a realistic plan using self-testing and spaced review. I will make the final priorities.

BEFORE YOU USE THE RESPONSE

A 60-second quality check.

Use these four questions before AI output becomes part of your notes, analysis or assignment.

01

Can I explain this without AI?

If the explanation falls apart, you have found the place where you need to keep working.

02

Have I seen the original source?

Check the name, year, quotation, concept and context in the source itself.

03

Are the academic choices mine?

You should be able to justify the problem, theory, method, analysis and conclusion.

04

Does this use follow the rules here?

Check the course description and exam instructions for the specific course and assessment.

CBS rules change over time

Use the official rules collection as your route to the current guidance. The course description and exam instructions tell you what applies in your specific situation.

FURTHER READING

The research behind this guide.

The studies suggest that both tool design and patterns of use shape learning. Results vary across subjects, tasks and students, so these links cover benefits, risks and established learning science.

01 · FIELD EXPERIMENT · AI TUTOR

Generative AI Without Guardrails Can Harm Learning

Bastani et al. · 2025

A large mathematics experiment in which a tutor with learning safeguards produced better outcomes than unrestricted access to a standard chat interface.

02 · EXPERIMENTS · DEPTH OF KNOWLEDGE

LLM syntheses and depth of learning

Melumad & Yun · PNAS Nexus · 2025

Seven experiments on how ready-made AI syntheses can reduce the active exploration and personal synthesis involved in developing deeper knowledge.

03 · RANDOMISED STUDY · HIGHER EDUCATION

AI Tutoring Enhances Student Learning

Fischer, Rau & Rilke · IZA · 2025

An experiment with university students that finds learning gains from AI tutoring and highlights the role of self-regulation and usage patterns.

04 · SYSTEMATIC REVIEW · HIGHER EDUCATION

The Impact of GenAI Chatbots on Student Learning

Badger, López & Nissen · 2026

A CBS review of 49 empirical studies covering engagement, comprehension, critical thinking, self-regulation and learning performance.

05 · CLASSIC STUDY · RETRIEVAL PRACTICE

Test-Enhanced Learning

Roediger & Karpicke · Psychological Science · 2006

The classic experiment behind self-testing: active retrieval improved retention over time more than repeated studying.

06 · RESEARCH REVIEW · STUDY TECHNIQUES

Improving Students’ Learning With Effective Learning Techniques

Dunlosky et al. · 2013

A broad review of ten study techniques, including practice testing, spaced practice, rereading and highlighting.