Sudden and rapid mass movements associated with landslides and snow avalanches present a hazard to life and infrastructure, yet their predictions and triggering mechanisms remain poorly understood. Statistical methods and correlative studies have been used to produce landslide or avalanche susceptibility maps, often with limited physical basis. Mechanistic approaches based on factor-of-safety computation seldom represent the progressive transition from local failure events to a landslide and, in general, do not include heterogeneities associated with land cover or with subsurface material properties and hydrologic pathways. Focusing on rainfall-induced shallow landslides, we propose the use of the fiber bundle model (FBM), a generic yet powerful and adaptable model used in modeling fatigue and fracture of complex and disordered materials. The primary strength of the FBM is its ability to represent the progressive failure of cracks and shear zones and the ruptures of highly heterogeneous bonding elements that are present in soils at all scales. The model also provides a natural framework for interpretation of acoustic emission signatures from failing slopes, which may form the basis of a monitoring and warning system.