GENERATIVE ARTIFICIAL INTELLIGENCE FOR EARTH SYSTEM

Outputs

Expected outputs of GenAI4Earth are the following :

  • GenAI-enabled FAIRification toolkit for AI-ready datasets
  • GenAI quality toolkit embedding bias detection, provenance, and trust metrics
  • Validated pilots with multimodal foundation models (urban, agro-forestry, seismology)
  • Machine-actionable APIs and reproducible workflows aligned with EOSC Nodes
  • EOSC-aligned GenAI training suite and onboarding resources
  • Policy briefs and standardisation contributions under GenAI4EU and EOSC

Specifically, this will mean :

  • 90% of project datasets reaching “GenAI-ready” FAIR level
  • ≥3 multimodal foundation models validated in pilots
  • ≥3 APIs/services integrated with EOSC infrastructures/nodes
  • ≥100 researchers trained and ≥50 data stewards onboarded
  • ≥6 policy contributions submitted to EOSC/EC bodies
  • Demonstrators delivered in pilots: UC1 Urban resilience, UC2 Agro-environmental monitoring, UC3 Hazard detection (land–ocean systems)