Mainstreaming Earth Intelligence: Harnessing Foundation Agents in Earth Observation

AI has already transformed Big Earth Data into actionable Earth Intelligence, enabling resilient agriculture and urban planning, enhancing the monitoring of ice dynamics and permafrost, tracking wildfires and biodiversity, and improving disaster recovery efforts. Today, the rapid advancement of frontier AI, including the rise of foundation agents, signals a profound evolution across the entire Earth Observation (EO) process and ecosystem. At the same time, EO must be mainstreamed as both a digital public good and digital public infrastructure, ensuring equitable access and enabling evidence-based decision-making where it matters most: at the front lines of sustainable development.

This 2025 DBAR workshop brings together EO practitioners and stakeholders to share practical experiences in applying AI across EO functions, including data acquisition, processing, analysis, and operational use. The workshop will highlight emerging opportunities and persistent challenges related to data quality, model scalability, associated risks, governance frameworks, and the importance of interdisciplinary collaboration.

The goal is to distil lessons from real-world AI applications and collaboratively develop a roadmap for the future of EO—one that fully leverages foundation agents alongside existing capabilities. Such a roadmap should not only demonstrate how AI can enhance perception and inference across scales, but also help close the usage gap for decision-makers working at the forefront of digital sustainability.