FAIRenrich: Scaling FAIR Semantic Annotation Workflows Through Modular AI Integration

The rapidly expanding volume of Earth System Sciences (ESS) data and persistent gaps in comprehensive semantic annotation present critical challenges for advancing interdisciplinary research and enabling robust scientific data reuse. FAIRenrich directly addresses these challenges with a core emphasis on scalability. Its open source, modular and highly configurable architecture is purpose built to efficiently scale semantic annotation workflows across institutional boundaries and handle diverse data volumes and complexities. By seamlessly integrating multiple NLP backends including spaCy, Ollama and GPT4All and utilizing the TIB Terminology Service for semantic enrichment, FAIRenrich adapts flexibly to a variety of institutional and disciplinary contexts, all without requiring code modifications. The framework empowers organizations to enrich large, heterogeneous and multilingual datasets using dynamic data pipelines, for example CSV files, Postgres databases or real time web annotation. With transparent review workflows, dynamic terminology selection and actionable analytics for quality assurance, FAIRenrich advances both performance and FAIR compliance, strengthening semantic connectivity and long term interoperability within the ESS and Biodiversity research communities. The tool's technical design emphasizes reproducibility and institutional scalability. A built-in caching mechanism optimizes performance, reducing operational costs and enabling deployment in resource-constrained environments. By combining automated terminology annotation with human-in-the-loop validation, FAIRenrich addresses the practical challenge of large-scale semantic enrichment while maintaining annotation quality and contextual accuracy. This poster demonstrates how FAIRenrich delivers scalable semantic annotation workflows by leveraging a modular AI integration architecture. The solution is specifically designed to address the challenges of rapidly growing and diverse Earth System Sciences data, enabling efficient, large-scale annotation across institutional and infrastructure boundaries. FAIRenrich’s flexible framework supports both automated terminology assignment and expert validation, ensuring high annotation quality while maintaining adaptability and resource efficiency. Caching mechanisms and configurable interfaces empower organizations to deploy FAIR-compliant enrichment processes at scale, transforming annotation from a bottleneck into an interoperable and reusable asset for open science.