Hydrometeorological hazards, including floods and droughts, are among the most severe and costly natural hazards globally. AI is believed to have the potential to better predict, detect, and monitor these hazards. High-quality AI training data are dependent upon in situ measurements for adjustment, assimilation, and calibration. Here we discuss the risk that AI progress will be severely impaired by increasingly insufficient access to in situ measurements and provide suggestions for countermeasures.

A recent study by Dandabathula et al. attributes the 5 August 2025 Dharali disaster to an ice-patch collapse based largely on satellite imagery. Here, we examine remote sensing and process-attribution uncertainties in that interpretation. The proposed mechanism by Dandabathula et al. lacks spectral validation, geomorphic consistency, volumetric support, and geophysical corroboration. Available independent observations instead indicate rainfall-triggered mobilisation of unconsolidated paraglacial sediments, underscoring the need for rigorous process attribution in Himalayan hazard assessment.

Impact-based Forecasts and Warnings (IbFWs) are crucial for disaster risk reduction; however, most systems focus on single hazards, overlooking the complex interactions of multi-risk events. This siloed approach can underestimate impacts, especially when hazards occur simultaneously or sequentially. Developing multi-risk IbFW systems requires interdisciplinary collaboration, improved modelling, more impact data and clearer terminology. Analysis of historical disasters shows that multi-hazard events cause disproportionate economic losses. Advancing IbFW systems demands pragmatic innovation, robust datasets, and inclusive strategies to better reflect real-world hazard complexity.

This Comment analyzes a global online landslide classification challenge to examine whether AI competitions can help participants engage with international standards. The results show that while technical innovation was strong, awareness of standards on transparency, sustainability, and bias was limited. The prototype used for this competition can serve as a scalable mechanism to foster the responsible use of AI for disaster risk reduction and beyond.