Séminaire interne du LERN – Jeudi 8 octobre à 12h en Salle C3-19 (Campus Pasteur)
Jeudi 8 octobre 2026, nous aurons le plaisir d’écouter Olivier Beaumais (LERN, Université de Rouen) qui présentera son papier « Public trust and perceived environmental impact of artificial intelligence » (co-écrit avec Dorothée Charlier, Olivier Gergaud, Rosanne Logeart).
Abstract:
Public trust in artificial intelligence (AI) is a crucial yet poorly understood determinant of AI adoption and its associated productivity gains. Drawing on a nationally representative French survey (n = 3,403) and using a copula-based approach, we jointly model individual trust in AI and perceived environmental impact of generative AI relative to standard web search. For a reference respondent, we identify a negative dependence between the two perceptions, showing that heightened environmental impact perception correlates with lower trust in AI. We show that conditional probabilities differ sharply: the probability of expressing full trust in AI is thirteen times higher when respondents perceive AI as environmentally benign than when they perceive it as harmful. These findings show that environmental perceptions and public trust in AI are closely intertwined, highlighting environmental communication and disclosure as potentially important dimensions of AI governance and societal acceptance.