In a preexposure phase, all the participants were instructed to conduct a task involving mental mathematical operations. For Group EXPOSED, the instructions for this task, and the operations to be performed, were introduced by the image of a robot. Group CONTROL was not exposed to the image of the robot during this phase. At the beginning of the subsequent change blindness phase, all the participants were exposed to the image of the robot for 60 s. During this time, they were informed that several changes were to be included in that image, and that their task was to detect the greatest number of possible changes. Immediately after, a sequence of eight 2-s. screens with the image of the robot was initiated, and a blank screen lasting 0.5 s. was interspersed between them. The original image of the robot was presented first, and 12 changes were progressively included on the following screens. The changes most easily detected in Group CONTROL were detected less readily in Group EXPOSED, and the changes least easily detected in Group CONTROL were more readily detected in Group EXPOSED. These results are explained as a consequence of a salience-reducing effect produced by preexposure that would first affect the initially most salient stimulus features, thereby increasing the relative salience of the initially least salient features.