4 questions for Emmanuelle Houet (EDHEC PiLab) on educational innovation and the role of artificial intelligence

What exactly is the PiLab?

The PiLab is the educational innovation centre of EDHEC Business School. Created in 2014-2015, on the initiative of professors keen to exchange ideas and share their (best) teaching practices as well as everyday challenges, the PiLab has its origins in what were then bimonthly lunches: ‘Food for pedagogical thought’.

Today, the PiLab comprises nine full-time staff, more than 50 online workshops every year to facilitate professors on all campuses whether permanent or part-time, countless initiatives to support professorsand programmes, and a constant quest to facilitate and improve the student and teaching experience.

Artificial intelligence is undoubtedly one of PiLab's major projects, isn't it?

The PiLab embraced the subject of AI very early on. In February 2023, the first information and awareness workshop on the challenges of AI in education was organised for professors.

And since then, the PiLab has focused its activity on three objectives:

- To foster excellence in the  mastery of the subject of AI in education, with a proactive approach to understand the technical and organisational issues as well as monitor developments in products emerging on the market.

- To disseminate the key issues concerning AI to professors, but also to the school management.

- To train and assist professors with AI tools and provide support, on request, concerning the appropriate pedagogical approaches that needed to be adopted and  to determine which tools should be discarded, favoured or even reactivated according to the disciplines taught.

All this is done in a two-pronged approach: the laboratory is both proactive and quick to respond to requests from professors, who want to test a new approach or a new tool that is not yet referenced.

Could you give a concrete example of an AI use project?

The exploratory work carried out by E. Deglaire and P. Daly (1) on the use of AI in marking exams has launched a major project at school level.

An initial test was carried out in 2023 on a small number of exams (around fifty), which had already been marked, and whose grades had already been communicated to the students. These archived exam papers were used to test the tools, without any issues of process or deadline. Our first conviction was that the professor must remain totally responsible for marking, and that the future does not lie in considering outsourcing the marking task to the tool. Rather, it is a question of learning to work intelligently with AI tools.

In the spring of 2024, on the strength of this promising first test, we carried out real-time tests on real exam papers. The professor had previously inputted her correction instructions and grading scale into the AI tool. She then went back to the AI to validate or edit the grade and/or the automatically generated comments. At this stage, no real time savings were observed because the implementation was still experimental, with internal processes not yet fully optimised, and there were too few students to take full advantage of the return on investment.

However, the information spread within the school and during the autumn of 2024, the roll-out continued with larger numbers of students and in other disciplines. A small group of professors supported by the PiLab, is currently embarking on the adventure to gain their own experience and expand the collective experience.

What is impressive is that EDHEC's culture of ‘learning by doing’ has enabled the PiLab to take this project from experimentation to theoretical analysis, to the development of a conceptual framework, and then to implementation in just a few months.

AI in correction assistance, like AI in education more broadly, raises a number of questions. In your opinion, what are the main issues at stake?

Indeed, there are many questions: will AI change the role of teachers? Or more specifically, here, will assisted correction relieve professors? Will it enable faster, more personalised feedback to students? Or will AI simply be confined to better quality control as a tool for double marking? Will it reduce teachers' marking time, allowing them to spend more time face to face with students? Will it improve the quality of individual monitoring?

We sense that AI-assisted marking of exam papers may strengthen the relationship between students and their professors or, on the contrary, be a step towards dehumanised teaching, depending on the organisational response that is put in place.

The PiLab has a duty to provide solutions and to act as a catalyst for structural reflection, without being too naïve and being totally pragmatic. That is why we have launched a study with teachers on the social acceptability of AI-assisted marking, using a large 360° questionnaire of all stakeholders within the school. This initiative is in line with our mission at the PiLab  to fuel the educational debate and help the school community to progress along the great and inevitable path of AI, with confidence and  clarity.

References

(1) AI-assisted grading: feedback on a full-scale test at EDHEC (March 2025) EDHEC Vox - https://www.edhec.edu/en/research-and-faculty/edhec-vox/ai-assisted-grading-feedback-full-scale-test-edhec-artificial-intelligence