When artificial intelligence is used for educational purposes

For a long time, creating a MOOC (Massive Open Online Course) meant much more than just designing videos and structuring coherent educational content for teachers. It also involved creating a large number of quizzes to accompany learners throughout the course: comprehension, practice, and assessment quizzes. A simple but demanding rule had been established: approximately one question for every two minutes of video.

This was a colossal task, often underestimated.

“It was undoubtedly one of the most time-consuming parts of creating the MOOC”, explains Sharon Tal-Itzkovitch, co-leader of a new course being developed alongside Marc Gruber.

Today, the Center for Digital Education has decided to support teachers. Using artificial intelligence, the team has implemented a semi-automated quiz generation process that significantly reduces the workload for professors.

About the importance of double-checking: behind the scenes of a discreet but powerful tool

Contrary to popular belief, AI does not work alone. Behind each series of quizzes generated lies meticulous design work carried out by a team of educational engineers and IT specialists. Together, they have designed a specific prompt that enables AI to generate questions tailored to the expected academic level, the subject matter, and the educational objectives of the course.

“The idea is not to replace the expertise of teachers, but to support it”, emphasizes Christian Vonarburg, Course development manager at the Centre for Digital Education. “AI acts as a first draft: it prepares the raw material, which teachers can then refine, correct, and adapt”.

In practice, the process follows several specific steps:

1. Quizzes are generated from the content of MOOC videos.
2. Quiz files (one per video) are formatted in a readable, straightforward, and easy-to-correct format and contain the video link, allowing teachers to quickly find the relevant passage.
3. The teacher then reviews the content: they can correct wording, add or delete questions, and adjust the level of difficulty.
4. Once validated, the files are sent back to the Center for Digital Education. The questions are then converted to XML format and automatically integrated into the course using a reference system based on the course index.

This process is invisible to the learner, but extremely effective for the teaching teams.

A concrete example: a MOOC on technological innovation and market opportunities

Sharon Tal-Itzkovitch and Marc Gruber are currently working on a new MOOC focusing on market opportunities generated by new technologies (From Lab to Market - Key steps to identify your most promising market opportunities). This is a challenging subject, at the crossroads of innovation, entrepreneurship, and strategy.

Such content requires quizzes that are up to the task: precise, conceptual, applied, and sometimes complex.

“The questions generated by AI focus on elements that I would have chosen myself: key definitions, process steps, concrete applications,” notes Sharon.

AI saves time... and offers new educational freedom

By automating a significant part of the process, the Center for Digital Education allows teachers to focus on what matters most: content quality, educational reflection, and interaction with learners.

“This use case for AI is interesting because it allows to support a core mechanisms of learning: feedback. Students benefit from being able to check their understanding. Of course, if they use AI to answer the questions, the effect will be zero. But, for those who are interested in learning something, having extra quizzes available is a good way to engage with the content.” adds Patrick Jermann, Executive Director of the Center for Digital Education.