We aim to develop and evaluate an AI-driven service that enables natural language–based findability of geospatial data within the Geoportti research infrastructure.
By studying how AI, and particularly large language models (LLMs), can improve FAIRification, the work explores metadata enrichment, robustness under noisy data, algorithmic consistency, and user adoption, while extending the vision toward AI-AIR (Accessibility, Interoperability, Reusability). The research addresses societal, economic, and environmental challenges linked to limited access and usability of geospatial data. Improved data findability supports evidence-based decision-making in areas such as climate change adaptation, land use planning, biodiversity monitoring, and sustainable urban development, while also reducing costs for data discovery and enabling broader participation by non-experts. Methodologically, the project combines systematic reviews, experimental studies with Geoportti datasets and adoption studies
