Artificial intelligence to advance Earth observation: : A review of models, recent trends, and pathways forward

AI is transforming how we monitor Earth's health: The paper emphasizes that artificial intelligence and machine learning are essential tools for converting massive amounts of raw satellite data into actionable information about land and ocean processes, enabling better planetary monitoring and decision-making.

Explainability and transparency matter for Earth observation: The review highlights "Explainable AI and causal inference" as a dedicated focus area, suggesting that for responsible deployment in environmental monitoring, AI systems need to be interpretable so scientists, policymakers, and the public can understand why the technology makes specific predictions about environmental changes.

Physics-based AI offers more trustworthy environmental predictions: The paper identifies "Physics-aware models" as a key pathway forward, indicating that combining AI with established scientific principles produces more reliable results than purely data-driven approaches alone, which is crucial when informing environmental policy.

Ethical and societal considerations cannot be an afterthought: The review dedicates a section to "ethical and societal issues related to the massive use of ML technologies in EO," signaling that responsible AI development in Earth observation requires proactive attention to these concerns rather than treating them as secondary considerations.

User needs should guide AI development in environmental monitoring: The inclusion of "User-centric approaches" as a core topic suggests that effective Earth observation AI must be designed with input from the scientists, governments, and communities who will actually use these tools to address real environmental challenges.