Reliable quizzes and answer keys

How can you use AI to obtain a quiz that faithfully reflects learning material, without ambiguous questions or invented answers?

1250 € per session, excl. VAT

In short

This half-day session of 3 hours offers a guided introduction to AI for creating a quiz from learning material. Participants write a prompt, examine the resulting questions and check the answer key. They leave with ten revised questions, their explanations and a completed checklist.

Why this training?

A quiz can seem convincing while testing secondary details or offering several defensible answers. Starting from knowledge explicitly present in the learning material makes it possible to check what each question measures and to justify the answer key. AI helps suggest wording and distractors, but may add a rule that is absent from the document. The training therefore prioritises comparison with the source, examination of ambiguities and human revision rather than immediate acceptance of the generated questions. Running case study: The fictitious organisation « Silvoret Médiation » welcomes volunteers to a community association's library. Using a fictitious information sheet specifying loan periods, renewals and the handling of returns, it wants to check ten useful items of knowledge before their first shift. Transfer to the workplace: After the session, reuse the prompt with another short piece of approved learning material. Have the quiz reviewed by someone familiar with the content, then note any questions misunderstood during its first use so that they can be reworded.

Objectives of this training

  • Identify knowledge in learning material that can be assessed.
  • Write a prompt specifying the audience, source, format and number of questions.
  • Produce ten single-answer multiple-choice questions with an explanatory answer key.
  • Verify each answer and its explanation against the source document.
  • Revise ambiguous wording and implausible distractors.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Occasional trainers wishing to check learning after a presentation.
  • Training officers and learning material designers who are new to AI.
  • Internal contact persons preparing knowledge questionnaires for an induction.

Prior skills

  • No prior AI knowledge required.
  • Ability to read short learning material and use a browser.

What you will work on

  1. Choosing what the quiz checks

    Objective: Distinguish verifiable knowledge from vague wording. The role of a knowledge quiz; the difference between memorisation and understanding; limitations of AI suggestions.

  2. Formulating a first guided request

    Objective: Obtain a draft constrained by the source. Structured prompt: beginner audience, ten questions, three options, a single correct answer, explanation and source reference. Ask for insufficient information to be flagged.

  3. Checking questions and answer keys

    Objective: Detect factual errors and ambiguities. Credible distractors, negatives, unintended clues, duplicate questions and invented information; systematic checking against the learning material.

  4. Testing and finalising the quiz

    Objective: Check the readability of the questionnaire before use. Learner instructions, a separate answer key, a simple marking scheme of one point per correct answer; distinction between quiz quality and the score achieved.

What you take away

  • Table of the ten targeted items of knowledge and their source passages.
  • Reusable prompt for requesting a quiz based on a document.
  • Ten-question quiz, with a learner version and a separate explanatory answer key.
  • Completed checklist documenting the corrections made.

Method and assessment

  • Short instructional inputs followed by immediate practice using the running case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Arrange access to a generative AI tool approved by the organisation. Prepare one to two pages of fictitious or anonymised learning material; the case study provided allows participation without a personal document.

Observable criteria: Each of the ten questions corresponds to an item of knowledge identified in the learning material. ; Each question has three options and only one demonstrably correct answer. ; Each answer key quotes the source passage and explains the correct answer. ; Distractors are plausible, with no grammatical clue or revealing difference in length. ; The checklist counts the questions meeting the criteria before and after revision and flags points still requiring validation..

Frequently asked questions

Do you need to have used AI before?

No. The first request is built step by step from the learning material provided.

Can we use our internal documents?

Only if they are anonymised and approved for use with the chosen tool. No individual learner results are needed.

Is the proposed answer key automatically reliable?

No. Each answer must be checked against the source; a question that cannot be verified is removed or reworded.

Will the quiz be integrated into a platform?

The session provides content ready to transfer. Importing it into or configuring a platform is not included.

Last updated: