The article examines the evolution of library reference services in the digital age, with a focus on the impact of generative artificial intelligence, particularly Large Language Models (LLMs). While initially prone to “bibliographic hallucinations,” these systems can improve through advanced prompting techniques, role assumption strategies, the use of reasoning models, and integration with traditional information retrieval tools. Through a practical example of bibliographic recommendation, the article demonstrates how different LLMs, when carefully selected, configured, and used with expertise, can generate accurate bibliographic suggestions by combining generative capabilities and information retrieval. While AI cannot replace the broad range of activities and competencies that define the library profession, it proves to be a valuable support tool for reference services, provided that innovation, methodological rigor, and ethical awareness are balanced.
L’articolo analizza l’evoluzione dei servizi di reference bibliotecari nell’era digitale, con specifico riferimento all’impatto dell’intelligenza artificiale generativa, in particolare i Large Language Models (LLM). Sebbene inizialmente soggetti ad 'allucinazioni bibliografiche', questi sistemi possono migliorare mediante tecniche avanzate di prompting, assunzione di ruoli specifici, modelli di ragionamento e integrazione con strumenti di ricerca tradizionali. L’articolo mostra, attraverso un esempio pratico di raccomandazione bibliografica, come LLM diversi, se scelti, configurati e utilizzati con competenza, possano generare raccomandazioni bibliografiche accurate unendo capacità generative e information retrieval. Pur non sostituendo l’ampio spettro di attività e competenze che caratterizza la professionalità bibliotecaria, l’IA si rivela uno strumento di supporto valido per il reference, a patto di bilanciare innovazione, rigore metodologico e consapevolezza etica.
Roncaglia, G. (2025). Verso il reference generativo?. BIBLIOTECHE OGGI TRENDS, 11(1), 63-73.
Verso il reference generativo?
Gino Roncaglia
2025-01-01
Abstract
The article examines the evolution of library reference services in the digital age, with a focus on the impact of generative artificial intelligence, particularly Large Language Models (LLMs). While initially prone to “bibliographic hallucinations,” these systems can improve through advanced prompting techniques, role assumption strategies, the use of reasoning models, and integration with traditional information retrieval tools. Through a practical example of bibliographic recommendation, the article demonstrates how different LLMs, when carefully selected, configured, and used with expertise, can generate accurate bibliographic suggestions by combining generative capabilities and information retrieval. While AI cannot replace the broad range of activities and competencies that define the library profession, it proves to be a valuable support tool for reference services, provided that innovation, methodological rigor, and ethical awareness are balanced.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


