AI-Powered Bibliographic Control: Automating Cataloging, Standardization, and Data Capture

Abstract

Artificial Intelligence (AI) optimizes bibliographic control by automating cataloging workflows, capturing data via mobile OCR, and normalizing records to established standards. Universal Bibliographic Control demands solutions that address the exponential growth of documentary output and the requirement for immediate access. AI accelerates record creation and review while maintaining consistency, even in resource-constrained environments. A human-in-the-loop system was implemented: librarians validate and correct AI-generated entries, reinforcing accountability and accuracy. Algorithmic transparency is ensured by documenting the AI’s criteria and decisions, creating an audit trail that facilitates suggestion evaluation and bias detection. Three use cases demonstrate the impact: 1. Automated Cataloging of Theses and Monographs. AI models segment PDF documents, extract essential fields, and generate MARC 21/RDA records. Average accuracy exceeding 93 % Cataloging time reduced from 30 minutes to under 5 minutes per item 2. Mobile Application for Image-Assisted Cataloging. Android devices capture cover pages; AI employs OCR and computer vision to propose normalized metadata. Provisional records are integrated into the bibliographic system via REST API. 3. Correction and Normalization of Historical Records. Models trained on RDA rules and authority lists detect inconsistencies and propose automatic or semi-automatic corrections. Over 40 000 records updated, 58 %...