Topic: Establishing an AI Practices Observatory in Higher Education: Current Trends and Future Directions

Lecturer: Dr. Pierre Bellet

Time: May 16, 2025, 9:00-12:00, UTC+8

Venue: Room 510, Xuehai Building, North Area of Nanhu Campus

Biography: Dr. Pierre Bellet is a postdoctoral researcher in Education and Training Sciences at Aix-Marscille University (AMU), where he coordinates the development of the Observatory of AI Practices in Higher Education within the DREAM-U project. His academic trajectory is grounded in a doctoral dissertation that explored the dynamics between collaborative efficacy and collective intelligence in digitally mediated learning environments-a theme that continues to shape his research into the educational potential and sociotechnical implications of artificial intelligence. Before transitioning into academia, Pierre spent nearly a decade as a creative technologist and product designer in the French Tech innovation ecosystem. This experience sharpened his expertise in user-centered design, digital interfaces, and collaborative tools-skills he later applied in higher education through pedagogical engineering roles. At the IUT de Montpellier-Sète, he co-designed immersive learning spaces, including a full-scale Learning Lab equipped with mobile technologies and XR tools to support active and blended learning experimentation. Now at AMU, his work spans instructional design, competency-based education, and digital infrastructure for student success. He actively contributes to multidisciplinary initiatives such as the Master’s in Pedagogical Engineering, the European UNI-T Academy, and AMU’s AI Observatory, which investigate how evolving technological ecosystems--including AI--can empower both learners and educators. Dr. Bellet’s research interrogates Al not merely as a tool, but as a catalyst for rethinking pedagogical relationships and technological practices. Drawing on principles of collective intelligence, he advocates for Al frameworks that are context-aware, co-constructed with practitioners, and aligned with the broader mission of higher education.

Abstract: As artificial intelligence becomes increasingly present in higher education, universities are confronted with the challenge of aligning institutional strategies with the complex realities of Al use in teaching and learning. This presentation introduces the ongoing development of an Observatory of AI Practices at Aix-Marseille University, aimed at documenting, analyzing, and supporting the evolving relationship between AI technologies and university teaching practices. We begin by surveying a series of grassroots and institutional initiatives, including tools like RaGaRenn or Lucie.chat, which exemplify the diversity of AI experiments taking place across national campuses. These projects serve as both pedagogical tools and real-time indicators of how faculty and students interact with AI in practice. Building on these observations, we propose a structured observatory model that combines technological monitoring, faculty development, and pedagogical research. This observatory functions as both a data collection platform and a collaborative space for exploring institutional responses--such as AI usage charters and localized initiatives underway in other French institutions. A key focus is the identification of mismatches between prescribed uses of Al(e.g., institutional guidelines or policies) and actual teaching practices. These gaps highlight the need for more adaptable and critically-informed strategies that can bridge policy with real-world adoption. Through mixed- methods research and field experimentation, the Observatory aims to develop tools and recommendations to inform institutional governance, faculty development, and ethical implementation frameworks. Ultimately, this initiative represents a forward-looking effort to move beyond reactive approaches to AI in higher education, by fostering shared understanding, critical reflection, and responsible integration at both strategic and classroom levels.

Rewritten by: Lin Qiaochu

Edited by: Liang Muwei, Li Tiantian

Source: School of Computer Science and Artificial Intelligence