Calls of several species of nonhuman animals are considered to be functionally referential. However, the cognitive mechanisms underlying audience behaviors remain unclear. This study used an audiovisual cross-modal preferential-looking paradigm to examine whether captive chimpanzees spontaneously associated a conspecific call with images of a corresponding item. Chimpanzees were presented with videos of snakes and fruit side-by-side while hearing playbacks of alarm calls, food-associated calls, or no sound (as a baseline condition). Chimpanzees looked at videos of snakes for longer when hearing alarm calls compared with food calls or baseline. However, chimpanzees did not look at videos of fruit for longer when hearing food calls compared with baseline. An additional experiment tested whether chimpanzees' gaze bias to the snake videos was driven by negative affective states in general via affect-driven attention biases. When chimpanzees were presented with the same snake and fruit videos while hearing playbacks of conspecific screams or no sound, they exhibited no gaze bias for snake videos. These results suggest that chimpanzees spontaneously associated alarm calls with images of a potential threat in a preferential-looking experiment and that this response was not simply driven by an affective state matching process. These findings should be interpreted in consideration of a procedural limitation related to pseudoreplication in the experimental stimuli. (PsycInfo Database Record (c) 2022 APA, all rights reserved).