https://www.nature.com/articles/s41561-025-01639-x

Dramsch, Jesper Sören ; M. Kuglitsch, Monique ; Fernández-Torres, Miguel-Ángel ; Toreti, Andrea ; Arif Albayrak, Rustem ; Nava, Lorenzo ; Ghaffarian, Saman ; Cheng, Ximeng ; Ma, Jackie ; Samek, Wojciech ; Venguswamy, Rudy ; Koul, Anirudh ; Muthuregunathan, Raghavan ; Hrast Essenfelder, Arthur

https://doi.org/10.1038/s41561-025-01639-x

This paper discusses the challenges and opportunities of using artificial intelligence (AI) in geoscience. While AI can analyse complex, multidimensional data and solve nonlinear problems, increasing model complexity often reduces interpretability. In critical contexts, such as natural hazard scenarios, this lack of understanding can undermine trust and limit implementation. The authors argue that explainable AI (XAI) methods, which make opaque ‘black-box’ models more interpretable, can enhance human understanding, build trust in model results, and promote broader adoption of AI in geoscience research and applications.

Information and views set out in this community page can also be those of the author and do not necessarily reflect the official opinion of the European Commission.