CAPSULE: Rethinking pedagogical and scientific practices in the era of generative AI…

Led by Sandrine Decamps and Bruno De Lièvre

Discover two complementary conferences to reflect on the challenges of generative AI (GenAI) in higher education and research. They provide key insights for integrating GenAI into your teaching and scientific practices in a critical, ethical, and effective way.

Please remember to bring your laptop: you will have the opportunity to test various tools and engage in hands-on activities.

Conference 1

How to Assess in the Age of AI? The ARIA Framework as a Pedagogical Compass

The rise of generative artificial intelligence in higher education is profoundly transforming assessment practices and challenging the very foundations of instructional design. This session stems from a research-based observation: educators are struggling to reposition their assessment methods in the face of tools that have spread much faster than the development of adapted pedagogical guidelines.

To address this challenge, Sandrine DECAMPS introduces the ARIA framework, structured around four complementary principles—Acceptability, Responsibility, Intelligence (Collective), and Accessibility—designed to guide the integration of AI into higher education. These principles aim to ensure that pedagogical uses remain legitimate, responsible, collective, and equitable, both in the design of training programs and in the transformation of assessment practices.

Conference 2

The Contribution of AI to Academic Research

This presentation explores how artificial intelligence can be leveraged to support academic research. After an introduction situating generative AI within the field's evolution—from symbolic AI to Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) systems—the session provides a structured overview of specialized tools available to researchers.

Six key functions of the research cycle are illustrated through dedicated tools:

• Literature search (ResearchRabbit, Consensus, Scite, Elicit)

• Information extraction (NotebookLM, ChatPDF)

• Idea organization (Connected Papers, Litmaps)

• Data processing

• Writing (Semantic Scholar, Scholarcy, SciSummary)

• Illustration and presentation (Gamma, Napkin)

This talk also offers insights into supporting and supervising the writing of master's theses and dissertations. The presentation concludes by addressing the ethical and deontological challenges associated with AI in research, drawing notably on the European Commission's guidelines for the responsible use of generative AI in scientific research.