Currently, many structures existing in seismic areas are highly vulnerable because they have been built without the use of seismic design codes or by using outdated codes. Often, methods for assessing the vulnerability of the structures do not take into account that their seismic behavior is dynamic and highly nonlinear and, moreover, that the structural characteristics and action have large uncertainties. This article aims to assess the vulnerability of structures taking into account that the mechanical properties of materials and the seismic action are random variables, by using advanced techniques based on the Monte Carlo method and on the nonlinear stochastic dynamics. The results obtained with these techniques are compared with those corresponding to a standard vulnerability assessment, based on deterministic models, in order to highlight the differences between both approaches. The main conclusion of this work is the need to address the vulnerability assessment problem from a probabilistic perspective which, combined with advanced nonlinear static and dynamic structural analysis techniques, provides a powerful tool giving information impossible to be captured by means of deterministic models. Finally, detailed results obtained for a building with waffle slabs, which is a structural typology widely used in Spain, are included and discussed.