Week 6: RUKF — Role, Task, Context, and Format

 


Welcome to Week 6 of the AI Literacy Course.

In Week 5, we learned that a useful first prompt begins with a purpose, gives AI a clear task, provides necessary context, requests a suitable response, and sets reasonable boundaries.

This week, we organise those choices using the RUKF framework.

RUKF helps turn a human intention into a clear instruction. It can make an AI response more relevant and easier to evaluate. However, it cannot guarantee that the response is correct, complete, fair, or suitable for its final use.

A note about the name RUKF

The abbreviation RUKF comes from the Swedish words:

  • R – Roll: Role

  • U – Uppgift: Task

  • K – Kontext: Context

  • F – Format: Format

In this English course, we use the words Role, Task, Context, and Format, while keeping the abbreviation RUKF.

Learning goals

By the end of this lesson, learners should be able to:

  • Explain the purpose of each RUKF element

  • Choose a role that describes useful behaviour

  • Write a specific and observable task

  • Provide relevant context using only necessary information

  • Select a format that supports learning and checking

  • Combine the four elements into one coherent prompt

  • Check whether an AI response followed the RUKF instruction



1. From purpose to RUKF

A prompt begins with a human purpose.

For example:

“I want to practise explaining inflation. I do not want AI to write a finished assignment for me.”

The purpose belongs to the learner. RUKF then turns that purpose into four practical questions:

ElementQuestion
RoleHow should AI behave during the task?
TaskWhat exact action should AI perform?
ContextWhat relevant information does AI need?
FormatHow should the response be presented?

The four elements should support the same purpose.

The categories are tools for thinking, not rigid boxes. Some instructions may fit more than one element. What matters is that the complete prompt is clear, coherent, and suitable for the task.

2. Role: How should AI behave?

Role describes the behaviour, perspective, tone, or working method you want AI to use.

Useful roles may include:

  • a patient study coach;

  • a critical reader;

  • an interviewer;

  • a plain-language explainer;

  • an organiser;

  • a discussion partner.

For example:

“Use the behaviour of a patient study coach. Ask questions before giving explanations.”

This tells the AI how to interact with the learner.

A role does not give AI real qualifications. Asking AI to “be my doctor,” “be my lawyer,” or “be my financial adviser” can encourage unsafe trust. The system may generate professional-sounding language without having responsibility for the consequences.

A safer instruction describes the limited behaviour needed:

“Explain this general medical term in plain language. Do not diagnose or recommend treatment.”

Role is not always necessary

Simple tasks may not need a role.

For example:

“Summarise the supplied paragraph in 80 words.”

Adding “Act as an expert summariser” may not improve this task.

Use a role when behaviour, tone, perspective, or interaction style matters. Do not add one merely to make the prompt longer.

3. Task: What should AI do?

Task describes the exact action AI should perform.

A clear task usually begins with an observable verb, such as:

  • explain;

  • compare;

  • classify;

  • summarise;

  • ask;

  • identify;

  • translate;

  • organise;

  • create practice questions;

  • give feedback.

Compare these examples:

“Help me with inflation.”

“Compare demand-pull and cost-push inflation using one everyday example of each.”

The second task is easier to understand and evaluate because the expected result is visible.

Weak tasks often use unclear expressions such as:

  • “Make it good.”

  • “Help me understand everything.”

  • “Make it interesting.”

  • “Tell me what I need.”

These instructions require the AI to guess what success means.

If a prompt contains several different tasks, divide the work into smaller stages. Begin with the first task and review its result before continuing.

4. Context: What does AI need to know?

Context provides the relevant background information that may change the response.

Useful context can include:

  • the intended audience;

  • the learner’s current knowledge;

  • the subject or situation;

  • the purpose of the task;

  • the language needed;

  • supplied source material;

  • important boundaries;

  • what has already been tried.

For example:

“I am an adult learner studying economics. I understand that prices rise during inflation, but I do not understand why different causes require different responses.”

This context helps the AI choose a suitable level and focus.

Use only necessary information

More context is not always better.

Irrelevant information can distract from the task. Personal or confidential information may create privacy risks.

Ask:

  • Does the AI need this detail to complete the task?

  • Could I describe the situation more generally?

  • Could I remove names and identifying details?

  • Could I use a fictional example?

Use enough information to make the task clear, but no more than necessary.

Removing a person’s name may not be enough if other details still reveal who the person is. Privacy and data minimisation will be explored more fully in Week 15.

5. Format: How should the response be presented?

Format describes how the response should be structured or delivered.

It may include:

  • headings;

  • bullet points;

  • a table;

  • a checklist;

  • a word limit;

  • a language level;

  • a number of examples;

  • step-by-step instructions;

  • one question at a time;

  • a short explanation followed by practice questions.

For example:

“Present the comparison in a two-column table. Finish with one short example and two questions that test my understanding.”

A useful format supports the learner’s next action. A comparison table can make differences visible. One question at a time can encourage active thinking. A checklist can support review.

However, format cannot guarantee truth. A clear table may still contain incorrect information.

Avoid conflicting requirements

Prompt instructions should be possible to follow.

For example:

“Give a very detailed explanation in no more than 30 words.”

The demand for extensive detail conflicts with the short word limit.

Too many headings, strict limits, or required items can also produce an unnatural response or remove important nuance.

Choose a format that fits the purpose rather than one that merely looks organised.

6. Make the four elements work together

A strong RUKF prompt is a coherent instruction, not four unrelated labels.

Consider this example:

Purpose:
I want to practise explaining photosynthesis rather than receive a finished answer.

Role:
Use the behaviour of a patient study coach.

Task:
Ask me five questions about photosynthesis.

Context:
I am an adult beginner. I sometimes confuse the roles of carbon dioxide and oxygen.

Format:
Ask one question at a time. Wait for my answer, then give brief feedback and one hint before continuing.

Combined prompt:

“Use the behaviour of a patient study coach. Ask me five questions about photosynthesis. I am an adult beginner and sometimes confuse the roles of carbon dioxide and oxygen. Ask one question at a time. Wait for my answer, then give brief feedback and one hint before continuing.”

The elements support one another:

  • The role creates patient, learning-focused behaviour.

  • The task defines what AI should do.

  • The context identifies the learner’s level and difficulty.

  • The format supports active participation.

The AI supports practice, but the learner still does the thinking.

7. Identify weak or missing elements

A prompt may contain all four labels and still be unclear.

Ask the following questions before using it:

Role

  • Does the role describe useful behaviour?

  • Does it falsely suggest professional authority?

  • Is a role necessary for this task?

Task

  • Is the action specific and observable?

  • Is there one main task?

  • Can I recognise whether it has been completed?

Context

  • Is the relevant audience or situation clear?

  • Have I included the information needed?

  • Have I removed unnecessary personal details?

Format

  • Will the structure support the next action?

  • Are the requested length and level reasonable?

  • Do any instructions conflict?

This check helps improve the prompt before testing it.

8. Test the complete prompt

After writing the RUKF prompt, test it with an appropriate AI tool.

At this stage, perform a basic RUKF alignment check:

  • Did the AI follow the requested role or behaviour?

  • Did it complete the stated task?

  • Did it use the relevant context?

  • Did it follow the requested format?

  • Did it add assumptions that were not requested?

A response may follow the format perfectly and still contain incorrect information. RUKF improves communication, not factual certainty.

Save the prompt and the first response. In Lesson 7, you will learn how to examine weaknesses, ask targeted follow-up questions, and improve the result.

9. RUKF is flexible

RUKF is a support for thinking, not a compulsory formula.

Not every prompt needs the same amount of detail. A simple task may require only a clear task and format. A complex learning activity may benefit from all four elements.

You can also write the elements in a natural paragraph. It is not necessary to include the labels “Role,” “Task,” “Context,” and “Format” every time.

The important questions are:

  • Is the purpose clear?

  • Does each instruction contribute to that purpose?

  • Is the task suitable for AI?

  • Can the output be evaluated?

  • Does the learner remain responsible for the final use?

Classroom activity: Build and test a RUKF prompt

Choose one current study problem.

Step 1: State your purpose

Complete this sentence:

“I want to use AI to help me…”

Be specific about what you want to understand, practise, organise, or create.

Step 2: Plan the four elements

RUKF elementYour choiceWhy is it useful?
Role
Task
Context
Format

If your task does not need a role, write “No role needed” and explain why.

Step 3: Combine the elements

Turn your choices into one natural prompt.

Check that:

  • the instructions support the same purpose;

  • the task is observable;

  • the context is relevant;

  • the format is realistic;

  • no unnecessary personal information is included.

Step 4: Test the prompt

Enter the prompt into an appropriate AI tool and save the first response.

Step 5: Complete the alignment check

Record whether the response followed:

  • Role

  • Task

  • Context

  • Format

Also note any assumption the AI added.

Step 6: Write a short reflection

In approximately 150–250 words, explain:

  • which RUKF element helped most;

  • which element was hardest to write;

  • whether the response followed the prompt;

  • what still needs to be checked or improved.

Save your prompt and response for Lesson 7.

Reflection questions

  1. Why does purpose come before RUKF?

  2. Why does the abbreviation use U for Task?

  3. When is a role useful?

  4. Why does a role not create professional qualifications?

  5. What makes a task observable?

  6. What context can improve a study response?

  7. Why should unnecessary personal information be removed?

  8. How can format support learning and checking?

  9. What can happen when prompt requirements conflict?

  10. Why can a strong RUKF prompt still produce an incorrect answer?

Key vocabulary

RUKF:
A prompting framework based on the Swedish words Roll, Uppgift, Kontext, and Format. In English: Role, Task, Context, and Format.

Role:
The behaviour, perspective, tone, or working method requested from the AI system.

Task:
The specific and observable action the AI should perform.

Context:
Relevant background information that helps the AI respond appropriately.

Format:
The requested structure, length, level, or delivery of the response.

Observable:
Clear enough that a person can inspect whether the task was completed.

Data minimisation:
Using only the information necessary for a task.

Alignment:
How well the AI response follows the role, task, context, and format in the prompt.

Coherent:
Clear and logically connected, with all parts supporting the same purpose.

Summary

In Week 6, we learned how to organise a prompt using RUKF:

  • Role describes how AI should behave.

  • Task states what AI should do.

  • Context provides the relevant information.

  • Format explains how the response should be presented.

A strong RUKF prompt connects all four elements to one human purpose. The role should not suggest qualifications the AI does not have. The task should be observable, the context should contain only necessary information, and the format should support the learner’s next action.

RUKF can improve clarity and relevance, but it cannot guarantee truth.


The framework is most useful when it helps learners express their purpose clearly while keeping understanding, checking, and responsibility in human hands.

Lesson 6 Interactive Quiz: RUKF

Choose one answer for each question. Then select Check my answers. You will receive feedback for every question. This practice quiz does not collect names or scores.

1. Why does RUKF use the letter U for Task?
2. Which question does the Role element answer?
3. Which role instruction is safest and most appropriate?
4. When might a prompt not need a Role?
5. Which Task is the most specific and observable?
6. Which Context is most appropriate for a study prompt?
7. A learner wants to compare machine learning and generative AI. Which Format best supports that purpose?
8. Which prompt contains conflicting instructions?
9. Which is the most coherent RUKF prompt?
10. An AI response follows the Role, Task, Context, and Format perfectly. What can the learner conclude?

Reflection: Choose one prompt you have written. Label its Role, Task, Context, and Format. Is any element unnecessary, missing, unclear, or in conflict with another element?