Accurate prediction of the occurrence of landslides triggered by rainfall can be challenging due to the uncertainties commonly associated with geotechnical, hydrological and climate parameters controlling the stability of slopes. Uncertainties may arise due to lack of knowledge (e.g., lack of geotechnical investigations) or the inherent natural variability (e.g., uncertain rainfall patterns) linked to these parameters. This work examines the potential of reducing uncertainties associated with rainfall-induced landslides by learning from observations of slope performances. The process of reducing the uncertainties is formalized by adopting the Bayesian updating framework. The performance of the implemented prediction models and the Bayesian updating framework is evaluated on the rainfall-induced landslides that occurred in 2011 in the Kvam area of central Norway. Slope failure or survival of a given rainfall event introduce additional information that can be utilized to reduce uncertainties associated with the assessment of rainfall-induced landslides in a certain area. Reducing uncertainties with site-specific observations of slope performance has a potential to increase the capacity of existing models to provide localized and more reliable predictions of rainfall-induced landslides.