AI is fundamentally changing the way we do science. However, making proper use of AI in Earth System Sciences (ESS) fundamentally relies on the ability to access and integrate heterogeneous observational, experimental and simulation data across the atmosphere, land surfaces, ocean, solid Earth, biodiversity and human-environment systems.
Research infrastructures have made major progress in aggregating, harmonizing, curating and providing access to these resources through FAIR data and interoperable services. The next challenge is to make this distributed knowledge ecosystem not only human-accessible, but also machine-actionable and AI-ready.
The rapid emergence of foundation models, generative AI and agentic systems is opening new opportunities to discover, integrate and analyse Earth System data, substantially improve physical modeling, automatize workflows and reach a new level of scientific knowledge discovery across disciplinary and organizational boundaries. Initiatives such as GenAI4Earth, EOSC-ARENA and Earth-AID illustrate this transition towards trustworthy and interoperable AI and generative AI tools embedded within the EOSC ecosystem.
This evolution raises a fundamental question: how can research infrastructures evolve from providers of FAIR data and digital services into an integrated ecosystem supporting AI-assisted and AI-driven scientific discovery?
The session will explore how research infrastructures, computing infrastructures and AI communities can work together to build this ecosystem. Contributions are invited that:
· demonstrate synergies between research infrastructures enabling cross-domain Earth System Science;
· present AI-ready data, generative AI services supporting scientific discovery;
· showcase integrated workflows combining data, models, and AI services across infrastructures;
· address interoperability, provenance, reproducibility, explainability, trustworthiness and European technological sovereignty;
· showcase approaches on how to build standards in ethics, scientific integrity or sustainability into AI services for ESS.
The session aims to bring together research infrastructures, AI developers and data communities with Earth system scientists and technology providers to discuss concrete pathways towards a federated European ecosystem in which FAIR data, AI models and computational services can be seamlessly combined to accelerate reproducible and trustworthy Earth System Science.