Research Summary: Instructors’ perceptions of a large language model framework for generating automated multiple-choice questions from video transcripts

A novel framework has been developed that leverages Large Language Models (LLMs) to automate the generation of multiple-choice questions (MCQs) from video transcripts. This framework specifically integrates an Item-Writing Flaws (IWF) checklist into the prompt engineering process, aiming to produce high-quality, criterion-aligned MCQs directly, rather than relying solely on post-hoc evaluation. This approach addresses the significant challenge instructors face in manually creating effective MCQs for video-based learning.

Why it matters

This development addresses operational inefficiencies in digital education by automating a time-consuming but critical task: the creation of assessment materials. It has the potential to scale quality content generation, directly impacting the effectiveness and accessibility of video-based learning and supporting consistent pedagogical standards across educational platforms.

Key insights

• Manual creation of high-quality Multiple-Choice Questions (MCQs) for video-based learning is a significant challenge for instructors.
• Video-Based Learning is identified as a central modality in digital education.
• Embedded MCQs are recognized as an effective strategy for promoting engagement and assessing comprehension.
• A new framework uses Large Language Models (LLMs) to automate MCQ generation from video transcripts.
• The framework explicitly integrates an Item-Writing Flaws (IWF) checklist into the LLM prompt generation, enhancing MCQ quality proactively.
• The process involves segmenting transcripts, extracting/classifying concepts, and applying structured prompting for generation.

Source

Educational Technology Research and Development — https://link.springer.com/article/10.1007/s11423-026-10646-7