A French ‘Grand École’ B-School Invests In AI To Rebuild The Classroom – Risks Included

IÉSEG School of Management is not just experimenting with AI in the classroom – it is betting on it.

The French business school has taken the unusual step of investing €500,000 in an AI startup founded by one of its own graduates, while committing roughly €2 million a year to embed the technology across teaching, operations, and student life.

“It’s not about technology for its own sake,” Dean Caroline Roussel tells Poets&Quants. “It’s about transforming how we teach and how students learn.”

THE BIGGER QUESTION: WHAT HAPPENS TO LEARNING?

The goal is straightforward: make AI a core part of how the school teaches, not just a tool students use on the side.

The risk is less so.

The tools are already in place. Students practice negotiation against AI-generated avatars. Finance concepts can be delivered through interactive AI instructors that can be paused, questioned, and revisited on demand. Faculty track student progress through dashboards that surface gaps in understanding in real time.

The promise is personalization at scale. But the concern, says Loïc Plé, director of pedagogy, is what that scale may erode.

“The potential erosion of students’ interest for learning – this is definitely a big risk,” Plé tells P&Q. “If we do not consider this risk, we are going to miss something.”

AI removes friction from the learning process. It also removes some of the struggle that often drives it.

OUTSOURCING THINKING – AND THE RISK OF DEPENDENCE

IÉ SEG Dean Caroline Roussel: “It’s not to invest in many companies. It’s to support our mission in teaching and learning”

A deeper concern is how students use the technology itself.

“This time, students can cognitively offload themselves on these AI tools,” Plé says.

Students can now generate explanations, draft answers, and solve problems with minimal effort. The question is no longer whether they can complete the task – but whether they understand it.

That uncertainty creates a second-order problem: assessment. If AI can produce high-quality outputs instantly, traditional ways of evaluating student work begin to break down.

“If you are not able to assess properly the competencies acquired by our students,” Plé says, “you can question the quality of the graduates that you put on the job market.”

CRITICAL THINKING WITHOUT FOUNDATIONS

Business schools have responded to AI by emphasizing critical thinking. Plé argues that may not be enough.

Critical thinking, he says, depends on underlying technical knowledge – the ability to judge whether an answer is correct, not just whether it sounds plausible.

“If they stop learning, how can they develop critical thinking?” he asks. “The quality of your critical thinking heavily relies on your level of technical skills.”

At the same time, companies are automating or eliminating entry-level tasks that once helped graduates build those skills. The combination raises a broader concern: students may be expected to evaluate AI outputs without ever fully developing the expertise to do so.

Plé’s concerns extend beyond pedagogy. “If students know that anyway an AI can do the job that they are currently learning,” he says, “then why should they study? Why should they work later?”

That question touches motivation – and, potentially, mental health. If the purpose of learning becomes unclear, the incentive to engage with it may weaken.

AN ELITE FRENCH SCHOOL MAKING AN UNUSUAL BET

Loïc Plé, director of pedagogy at IÉSEG: “If students know that anyway an AI can do the job that they are currently learning, then why should they study? Why should they work later?”

IÉSEG operates within France’s grande école system – a highly selective track of institutions that sit alongside public universities and have long served as pipelines to leadership roles in business and government.

Rather than limiting itself to partnerships, IÉSEG has taken an equity stake in Compleducation, an AI startup founded by one of its alumni and developed within the school’s incubator. The platform allows students to interact with AI “avatars” that deliver course content and respond to questions in real time.

The investment, Roussel says, is deliberate. “It’s not to invest in many companies,” she says. “It’s to support our mission in teaching and learning.”

A STRATEGY BUILT AROUND – AND AGAINST – THE RISKS

IÉSEG’s AI push is structured across six areas, including teaching and learning, research, administrative efficiency, stakeholder engagement, responsible AI development, and data governance.

A steering committee oversees the effort. A formal charter defines how AI can be used across the institution. Staff have been trained on internal tools designed to automate routine work while preserving human interaction.

In the classroom, the design reflects some of the risks Plé outlines. AI is used primarily to deliver foundational content and support students outside class. In-person time is increasingly reserved for discussion, debate, and applied work with companies – settings where students are expected to demonstrate understanding, not generate it.

Early use cases have focused on courses where students struggle most, with success measured in part by whether fewer students need to retake them. The approach is not to replace teaching, Roussel says, but to redistribute it.

THE BET

IÉSEG is moving faster than most business schools – investing in the tools, integrating them into courses, and scaling them across programs. 

It is also acknowledging, more directly than many peers, what those tools may disrupt. The same systems that promise more personalized, accessible learning also raise questions about motivation, mastery, and the role of human effort in education.

Plé does not dismiss those concerns. He treats them as part of the strategy.

“If we do not consider these risks,” he says, “we are going to miss something.”

DON’T MISS INSIDE THE BOLD TRANSFORMATION OF A ‘GRAND ÉCOLE’ BUSINESS SCHOOL

© Copyright 2026 Poets & Quants. All rights reserved. This article may not be republished, rewritten or otherwise distributed without written permission. To reprint or license this article or any content from Poets & Quants, please submit your request HERE.