Artificial Intelligence (AI) is transforming research workflows, data management, and software engineering across disciplines. This transformation brings both challenges and opportunities: how can AI be strategically integrated into research software development to enhance innovation, collaboration, and sustainability? And what new skills and expertise does AI-supported research require from research software engineers and providers of research services? This talk presents how the Software and Service Development Team at the Göttingen State and University Library systematically embeds AI within its developer pool and service portfolio. Our approach focuses on identifying AI competencies, fostering cross-project synergies, and establishing AI as a core capability of research infrastructure. We showcase practical applications across ongoing projects: The Artus Digital project uses AI-driven OCR pipelines for text segmentation, dewarping, and semantic structuring via large language models. In the digital edition project Ahiqar, AI enables precise line segmentation to align historical documents with digital transcriptions. OPERANDI applies machine-learning models to predict runtime and HPC resource requirements for large-scale digitisation. To foster AI expertise across the team, LibraryAI brings together HPC developers, doctoral researchers from the GippLab (Chair of Scientific Information Analytics, University of Göttingen), classical software engineers, and experienced research software engineers in an agile, PO-led team. This collaboration strengthens AI-related RSE capacity, accelerates innovation, and provides skill-development opportunities for all participants. By sharing our experiences, challenges, and governance insights, we aim to contribute to the community discussion on how AI can be responsibly and sustainably integrated into interdisciplinary RSE.
Schima-Voigt et al. (Thu,) studied this question.