Risorse archivistiche in rete nell'era dell'Intelligenza artificiale: ricerca, accesso e uso

This article examines how Artificial Intelligence is transforming the research, access, and use of archival resources within digital environments. Focusing on technologies such as Large Language Models (LLMs), semantic search systems, and Retrieval-Augmented Generation (RAG), the contribution analyses both the opportunities and the critical issues emerging from AI-driven archival mediation. While these tools can enhance access, discovery, and interaction through natural language interfaces and automated inference, they also introduce significant challenges related to transparency, reliability, contextual integrity, and algorithmic bias. Drawing on the InterPARES Trust AI project, the article discusses users’ perceptions of AI in archival access and highlights the need for responsible AI frameworks grounded in explainability, paradata, and archival principles. Finally, it argues that archival access should be understood not as a single retrieval action, but as a continuous and dialogic process shaped by the interaction between users, digital systems, and documentary contexts.