The brain combines information from different senses to improve performance on perceptual tasks. For instance, auditory processing is enhanced by the mere fact that a visual input is processed simultaneously. However, the sensory processing of one modality is itself subject to diverse influences. Namely, perceptual processing depends on the degree to which a stimulus is predicted. The present study investigated the extent to which the influence of one processing pathway on another pathway depends on whether or not the stimulation in this pathway is predicted. We used an action-effect paradigm to vary the match between incoming and predicted visual stimulation. Participants triggered a bimodal stimulus composed of a Gabor and a tone. The Gabor was either congruent or incongruent compared to an action-effect association that participants learned in an acquisition phase.We tested the influence of action-effect congruency on the loudness perception of the tone. We observed that an incongruent-task-irrelevant Gabor stimulus increases participant’s sensitivity to loudness discrimination. An identical result was obtained for a second condition in which the visual stimulus was predicted by a cue instead of an action. Our results suggest that prediction error is a driving factor of the crossmodal interplay between vision and audition.