This presentation examines the synergy between Linked Data, Artificial Intelligence (AI), and Large Language Models (LLMs) to enhance library discovery and services. We will explore methodologies such as Retrieval Augmented Generation (RAG) and the integration of linked data within LLMs for tasks like named entity recognition. In the context of knowledge graphs, we will demonstrate how these approaches improve AI output accuracy, ultimately enhancing user experience. Additionally, we will discuss questions posed by the LOD COP Working Group to Ex Libris to better understand our needs and inform their strategy for integrating AI and linked data solutions. Ethical considerations, including transparency in AI decision-making, user privacy, and equitable access to information, will also be addressed. Lastly, we will present ongoing experiments and innovative ideas that highlight the promising potential of combining AI and linked data in library discovery and metadata services, while promoting trusted and responsible AI applications.