Creating Exam Questions

Designing exams and quizzes can be time-consuming. AI can help by generating pools of questions or even full quizzes based on the topics you specify. For example, you might prompt, “Create 5 multiple-choice questions and answers to test understanding of [Topic X],” or “Propose two essay questions that require critical thinking about [Topic Y].”

The more detail you give the AI about the learning objectives or difficulty level, the better it can tailor the questions – for instance, indicating that you want application-level questions rather than factual recall. AI can even generate a rubric or scoring guide if you ask for it.

Again, treat these outputs as drafts: review each suggested question for accuracy and clarity, and ensure the difficulty is appropriate for your students. By iterating with the AI (e.g., “make the previous question harder” or “provide the answer key”), you can quickly develop a well-rounded set of assessments.

Practical Guide

How to prompt: Provide the topic, the type of assignment, and any specific angle or criteria you have in mind. If you need a certain format (e.g., multiple-choice vs. open-ended), make that clear.

1. Brainstorm ideas: If you’re not sure what assignment to give, ask the AI for ideas. Example Prompt: “I am teaching a history course on the Cold War. Suggest three creative assignment ideas that go beyond standard essays, possibly involving primary source analysis or role-play.” The AI might propose, say, a diary from the perspective of a historical figure, a debate simulation, or an analysis of declassified documents. These can spark your creativity and you can choose one to develop further.
2. Draft the assignment prompt: Once you have a direction, have ChatGPT flesh it out. Example Prompt: “Draft a detailed assignment prompt for a 2,000-word essay on Cold War history. The prompt should ask students to compare the foreign policies of two specific presidents, and include guidelines about using at least 5 scholarly sources and proper citations.” The model will output a paragraph or two that frames the task (e.g., “Analyze the approaches of President X and President Y…”) along with any criteria you mentioned. It may even add generic advice (which you can keep or trim).
3. Specify any structure or rubric hints: If you want the prompt to include how it will be evaluated or the key points to cover, you can ask for that too: “Add a bulleted list of what a successful essay should include (key considerations or sections).” This might produce something akin to a brief rubric or checklist within the assignment description, which can guide students.
4. Create variations or quiz questions (optional): For exams, you might generate multiple-choice questions or short-answer questions. Example Prompt: “Generate 5 multiple-choice questions (with answer keys) to test students on key events of the Cold War.” or “Give 3 short-answer exam questions focusing on causes of the Cold War, requiring a 2-3 sentence answer each.” . The AI will produce questions and likely the answers (which is helpful for your answer key). Review each question to ensure correctness and appropriate difficulty. You might need to fix any subtle inaccuracies in how options are phrased.

When using AI for assessment content, always ensure alignment with your learning objectives. To increase alignment of the output, include your learning objectives in the prompt. The AI might generate a question that is interesting but tangential; modify or discard those. Also consider whether an AI-generated prompt could be too formulaic – you might personalize it to your class context so it isn’t something easily found online. Many educators find AI useful for drafting assignment text which they then humanize and align to their course. This speeds up the process while keeping the instructor in control of the academic rigor.