Applications of artificial intelligence in audiovisual collections

metadata standards and language models in information organization

Abstract

Objective: This article presents a bibliographic review to examine the possibilities and challenges of using Artificial Intelligence (AI), with an emphasis on Large Language Models (LLMs), for the preservation and organization of audiovisual collections in Library and Information Science. Method: The literature review discusses digital preservation fundamentals, information organization, and the specific characteristics of audiovisual collections, linking them to AI concepts, prompt engineering, levels of control, and the methodologies of Chain-of-Thought (CoT) and Chain-of-Verification (CoVe). Results: Findings indicate that AI can automate tasks such as semantic indexing, transcription, and metadata generation, thereby increasing efficiency and access. However, issues such as data governance, algorithmic bias, implementation costs, and professional training remain significant challenges. Conclusions: The study concludes that AI does not replace librarianship but reshapes its role, requiring continuous education, strategic planning, and public policies to ensure high-quality and sustainable preservation of audiovisual heritage.