The impact of anxiety-provoking stimuli on the Sustained Attention to Response Task (SART; Robertson, Manly, Andrade, Baddeley, & Yiend, 1997), and response inhibition more generally, is currently unclear. Participants completed four SARTs embedded with picture stimuli of two levels of emotion (negative or neutral) and two levels of task-relevance (predictive or non-predictive of imminent No-Go stimuli). Negative pictures had a small but detectable adverse effect on performance regardless of their task-relevance. Overall, response times and rates of commission errors were more dependent upon the predictive value (relevance) of the pictures than their attention-capturing nature (i.e., negative valence). The findings raise doubt over whether anxiety improves response inhibition, and also lend support to a response strategy perspective of SART performance, as opposed to a mindlessness or mind-wandering explanation.