Send a class video to an AI-generated in-class quiz workflow to turn each lesson into a measurable learning check. Students can spot blind spots quickly, jump back to the relevant time code, and review with purpose. Teachers can identify common wrong-answer patterns, adjust pace and difficulty, arrange tiered support, and build a consistent question bank with less preparation and grading effort.
Below are suggested prompts and workflows for generating in-class quizzes with AI tools such as ChatGPT and NotebookLM . You can adjust and improve them for your own teaching needs.
Generate an In-Class Quiz with NotebookLM
1. Add the source transcript
Create a notebook in NotebookLM, then use Add source to upload the transcript file generated by Smartape recording system , Montage editing , or iFS self-edit upload .
2. Generate the question sheet without answers
Enter the following prompt in NotebookLM:
First, fully clean the uploaded transcript. Remove filler words, repetition, noise markers, and obvious typos while keeping the actual course content. Then generate a 20-question mixed in-class quiz question sheet that includes:
- About 5 true/false questions
- About 10 multiple-choice questions with four options, A-D
- About 5 fill-in-the-blank questions
Rules:
- Mix the question order randomly, cover important course concepts, and avoid repeating the same focus.
- For each multiple-choice question, list options A, B, C, and D below the question.
- Output only the question sheet in this step. Do not include answers or explanations.
- Use a clear list format so it can be copied into Word or Google Docs.
3. Generate the answer key
After checking the question sheet, enter the following prompt to produce the answer key:
Based on the previous question sheet, output the answer key in order.
Format:
- List answers by question number, 1-20.
- For true/false questions, mark True or False. For multiple-choice questions, list the correct option letter and add one short reason if needed. For fill-in-the-blank questions, provide the correct word or phrase.
- Output only the answer key. Do not repeat the questions.
4. Copy and export
Copy the question sheet and answer key into Google Docs or Word, then export as PDF for student use or archiving.
Generate an In-Class Quiz with ChatGPT
1. Upload the transcript
In a ChatGPT model that can handle long text, paste or upload the SRT transcript with complete time codes generated by Smartape recording system , Montage editing , or iFS self-edit upload .
2. Generate the question sheet and answer key
Paste the following prompt into ChatGPT:
Act as a teaching assistant. First, fully clean the transcript by removing filler words, repetition, noise markers, and obvious typos while preserving the actual course content. Then generate a 20-question mixed in-class quiz question sheet. Requirements: about 5 true/false questions, about 10 multiple-choice questions with four options A-D, and about 5 fill-in-the-blank questions. Cover the key concepts and facts from the course, avoid repeated question focuses, mix the order randomly, and keep the difficulty at a medium level. For each multiple-choice question, list options A, B, C, and D below the question. Output the question sheet without any answers or explanations, and also output a separate answer key. Use a numbered list from 1 to 20. Output in canvas format.
4. Copy and export
Copy both sheets into Google Docs or Word, format as needed, then save as PDF for publishing or printing.
Notice
The AI prompt guidance on this page is provided for reference and learning purposes only. Because AI technology continues to evolve quickly, some content may differ from the latest model versions or features.
Disclaimer
- BlueEyes Technology does not guarantee that the content on this page is always current, complete, or fully accurate.
- Users should evaluate and test prompt results independently, then adjust them according to actual needs.
- BlueEyes Technology is not responsible for any action taken based on this page or any resulting consequences, including but not limited to data accuracy, suitability, or direct or indirect loss.