Applying artificial intelligence technologies for acquiring digital collections in institutional repositories (Case study of “105 – ∞” collection in BNTU Repository)

Abstract

The article summarizes the practical experience of the Scientific Library of the Belarusian National Technical University (BNTU) in building a full-featured digital collection within the institutional repository with generative artificial intelligence (AI). The relevance of the study is determined by the need to master new technologies and to expand the range of library services and acquiring the unique digital content. The authors analyze and detail into the proven technique of building the digital illustrated collection «105 – ∞», applied in the BNTU Repository. The presented methodology covers all stages of the project: from concept development and selection of AI tools (ChatGPT, DALL-E 3) to generation of textual and visual content, document formation and allocation in the repository. The focus is made on the practical aspects of prompt engineering, metadata management, as well as ethical and copyright issues in the context of placing AIgenerated materials. The result of the study is not only the acquired collection, but also the scalable methodology, demonstrating the key advantages (automation, creation of innovative content) and limitations (the need for verification, adherence to ethical standards) of using AI. The accomplished case study is of practical significance, as other libraries and repositories can use its findings.