We find ourselves in a position where the near-exponential growth of our archive is causing several problems: our archive grows faster than our community's ability to navigate it, and faster than our ability to manage it. To combat this, on the user side, researchers increasingly work through AI-assisted tools. If NASA data is not findable and usable in those environments, we cede our role as the authoritative source for Earth science data. On the curation side, NASA needs to increasingly rely on automation. We will present our AI strategies targeting four areas: Production (AI-powered pipeline segments, mission data and metadata development); Infrastructure (open machine-readable interfaces, intelligent user support triage); Access (dataset matchmaking via external AI assistants, semantic discovery as a service); Analysis (Earth science notebook extension in JupyterLab, reproducibility scaffolding).
AudienceAnyone who wants or needs to employ AI/ML processes in their work
Agenda- Session Introduction - Doug Newman
- Earthdata MCP - Trevor Lang
- Earthdata AI Operations - Clement Li
- Earthdata AI Update & Plans - Cory Krause
- Agentic Queryable Earth: Conversational, Traceable Analysis Across Earth Science Data - Jason Gilman
- Surfing the Digital Wake - Handling Changing Data Access Patterns - Han Mai
- Cloud Cost/Budget Efficiency - Oholiab Gessesse
