LifeWatch ERIC as Catalyst and Connector: Scaling FAIR, AI-Ready Data Across Biodiversity, Ecosystems, and Climate Adaptation within EOSC Earth System Science increasingly requires seamless integration across disciplines and domains. Leveraging AI and generative AI demands data that is interoperable in machine-actionable ways. Yet making data truly AI-ready, structured for automated discovery, integration, analysis and reuse, requires coordination across research infrastructures. LifeWatch ERIC addresses this challenge by acting not as a standalone data repository, but as a catalyst for innovation and a connector across research infrastructures. In this role, LifeWatch ERIC provides analytical, semantic, and workflow-level bridges that enable data, services, and knowledge from infrastructures such as DiSSCo, eLTER, EMBRC-ERIC, AnaEE-ERIC and DANUBIUS ERIC, as well as global aggregators like GBIF, EMODnet and OBIS, to be combined into coherent, science- and policy-relevant networks. Concretely, LifeWatch ERIC provides a computational and semantic integration layer that turns distributed datasets and services into reusable workflows aligned with the EOSC Interoperability. We enable cross-RI composition through shared APIs, provenance-aware processing, and machine-readable descriptions of variables and methods, so that the same analytical logic can be executed across countries, domains, and observation systems. This integrative role is realised through several ongoing initiatives: (a) within ENVRI-Hub NEXT, LifeWatch ERIC collaborates with Data Terra, ICOS, ACTRIS, and other environmental RIs to deliver interdisciplinary services through the emerging ENVRI EOSC thematic Node, directly addressing cross-compartment data integration for environmental research; (b) through FAIR2Adapt (coordinated by LifeWatch ERIC), we are developing a FAIRification Framework for creating FAIR Digital Objects, demonstrated through six case studies spanning coastal ecosystem modelling in the Bay of Biscay, urban climate risk assessment in Hamburg, and national climate change adaptation hub development; (c) within EOSC Beyond, where LifeWatch ERIC is one of the 10 pilot nodes, we show how we can jointly support research communities thanks to the integration and interoperability between EOSC and Data Spaces by exploiting federating capabilities, and (d) through OSTrails, LifeWatch ERIC contributes to the design and piloting of end-to-end Plan–Track–Assess pathways, linking machine-actionable DMPs, Scientific Knowledge Graphs and FAIR assessment services, and demonstrating how environmental research infrastructures can operationalise FAIR-by-design workflows within EOSC. We present concrete approaches to AI-readiness, grounded in existing research practice: Discrete Global Grid Systems (DGGS) forproviding analysis-ready, multi-resolution data structures that unify heterogeneous sources into AI-accessible formats; AI-assisted metadata population reducing manual curation burden; and semantic interoperability through I-ADOPT, structuring variables into machine-readable components that enable cross-dataset discovery regardless of naming conventions. Rather than positioning AI as an end in itself, these demonstrate how research infrastructures can jointly shape EOSC for transnational, cross-domain challenges. To support trustworthy AI applications, we capture data licensing, provenance, quality signals, and uncertainty as first-class, machine-actionable metadata, including transparent records of when generative AI has contributed to metadata enrichment and whether human validation has been applied. This ensures that automation accelerates curation without weakening scientific accountability.
Fouilloux et al. (Wed,) studied this question.