GenAI Did Not Break Academia, It Revealed It: A Proposal for Structural Reform and Responsible AI Integration

Lobby D

Adam DEBBAH, CentraleSupélec Malo USELLE, CentraleSupélec Mahdi AYADI, CentraleSupélec

Daily life / student life / campus

We aim to redesign academic assessments and curricula to address the rise of generative AI and we want to properly integrate a controlled use of AI into the academic world with all the stakeholders(teachers, students and academic institutions). The project focuses on shifting from traditional "homework" to oral presentations and interactive defense of knowledge. This involves a deliberate simulation of a classroom where AI is integrated into the research phase, but the evaluation focuses on "raw sciences" and fundamental reasoning. Students and faculty collaborate to ensure that AI serves as a tool for deepening, rather than replacing, core intellectual labor. Controlled use of AI can be the usage of AI with a watermark AI model (with a private key shared between the ministry of education and that AI companies) so that AI can be used but the usage of it is assumed, AI to help schools and students to make better choices of courses and curricilums. Promoting the use of AI as teachers would be a first step and developing tailor-made AI teachers for homework would be great (with teacher monitoring the usage).

Current assessment models are struggling with the ubiquity of AI, leading to a potential decline in foundational knowledge, what we call "raw sciences." This situation creates a tension between efficiency and deep cognitive engagement. If students rely solely on AI for outputs, they risk losing the ability to perform first-principles thinking. This project is crucial for maintaining academic integrity and equity, ensuring that the "spirit of the curriculum" evolves to foster critical human intelligence and sustainability in learning rather than mere automation.

Our team is motivated by the desire to rehabilitate "pure" scientific reasoning in an AI-saturated world. We want to move beyond the binary of "banning" or "fully outsourcing" tasks to AI. Through this challenge, we hope to understand how to transform educational expectations to better reflect human cognitive strengths. Our goal is to question the current drift toward intellectual passivity and develop a framework where technology strengthens, rather than erodes, the student's mastery of fundamental concepts.

Please find our initial contribution in the following link: https://drive.google.com/file/d/1QJbU1ln1xTvMRMskOIirvLO2dEIQhtpl/view?usp=sharing