The “animate monitoring” hypothesis proposes that humans are evolutionarily predisposed to recruit attention toward animals. Support for this has repeatedly been obtained through the change detection paradigm where animals are detected faster than artifacts. The present study shows that the advantage for animals does not stand up to more rigorous experimental controls. Experiment 1 used artificially generated change detection scenes and counterbalanced identical target objects across two sets of scenes. Results showed that detection performance is determined more by the surrounding scene than semantic category. Experiment 2 used photographs from the original studies and replaced the target animals with artifacts in the exact same locations, such that the surrounding scene was kept constant while manipulating the target category. Results replicated the original studies when photos were not manipulated but agreed with the findings of our first experiment in that the advantage shifted to the artifacts when object categories replaced each other in the original scenes. A third experiment used inverted and blurred images so as to disrupt high-level perception but failed to erase the advantage for animals. Hence, the present set of results questions whether the supposed attentional advantage for animals can be supported by evidence from the change detection paradigm.