Bio:
Chloé Clavel is a Senior Researcher in the ALMAnaCH team at Inria Paris, the French national research institute for digital science and technology. Her research interests lie in the areas of Affective Computing and Artificial Intelligence, at the crossroads of multiple disciplines including speech and natural language processing, machine learning, and social robotics. She works on computational models of socio-emotional behaviors — such as sentiment, social stance, engagement, and trust — in both human–human interactions (e.g., conversations in social networks or face-to-face settings) and human–agent interactions (e.g., conversational agents and social robots).
Title : Understanding Social Interactions in the Era of LLMs – the Challenges of Transparency
Abstract
Research on AI and social interaction is not entirely new — it falls within the field of social and affective computing, which emerged in the late 1990s. To understand social interactions, the research community has long drawn on both artificial intelligence and social science. In recent years, however, the field has shifted toward a dominant focus on generative large language models (LLMs). These models are undeniably powerful but often opaque. In this talk, I will present our current work on developing machine learning approaches — from classical methods to LLMs — for modeling the socio-emotional layer of interaction, with a particular focus on improving model transparency. I will also briefly present some of the applications we are developing to support human skill development, particularly in the fields of education and health.