In nonhumans, ‘optimism’ is often defined as responding to an ambiguous item in the same manner as to items previously associated with reward (or lack of punishment), and “pessimism” is defined as responding to an ambiguous item in the same manner as to items previously associated with a lack of reward (or with punishment). We measured the degree of “optimism” and “pessimism” in three captive male western lowland gorillas (Gorilla gorilla gorilla) during four consecutive two-week periods in which the amount of available forage material (mulberry, Moraceae or alfalfa, Medicago sativa) was manipulated. We assessed cognitive bias using an ambiguous cue paradigm for the first time. Pairs of two-dimensional shapes were presented on a touch-screen computer in a forced choice task in which one shape was always reinforced (P), one was never reinforced (N), and one was reinforced half the time, making it ambiguous (A). The gorillas were presented with an equivalent number of PA and NA pairs prior to testing, which also included probe trials of ambiguous items paired with novel items. During the limited forage phase, the gorillas, as a group, selected the ambiguous stimulus (indicating optimism) at a level greater than chance; tests for all other phases were non-significant. The gorillas displayed individual differences in learning PA and NA trials and in their choice of ambiguous items on test trials. Idiosyncratic preferences for particular stimuli suggest that the ambiguous cue paradigm may not be ideal for assessing cognitive bias in nonhumans.