Reward-related stimuli can potently influence behaviour; for example, exposure to drug-paired cues can trigger drug use and relapse in people with addictions. Psychological mechanisms that generate such outcomes likely include cue-induced cravings and attentional biases. Recent animal data suggest another candidate mechanism: reward-paired cues can enhance risky decision making, yet whether this translates to humans is unknown. Here, we examined whether sensory reward-paired cues alter decision making under uncertainty and risk, as measured respectively by the Iowa Gambling Task and a two-choice lottery task. In the cued version of both tasks, gain feedback was augmented with reward-concurrent audiovisual stimuli. Healthy human volunteers (53 males, 78 females) performed each task once, one with and the other without cues (cued IGT/uncued VGT: n = 63; uncued IGT/cued VGT: n = 68), with concurrent eye-tracking. Reward-paired cues did not affect choice on the Iowa Gambling Task. On the two-choice lottery task, the cued group displayed riskier choice and reduced sensitivity to probability information. The cued condition was associated with reduced eye fixations on probability information shown on the screen and greater pupil dilation related to decision and reward anticipation. This pupil effect was unrelated to the risk-promoting effects of cues: the degree of pupil dilation for risky versus risk-averse choices did not differ as a function of cues. Taken together, our data show that sensory reward cues can promote riskier decisions and have additional and distinct effects on arousal.SIGNIFICANCE STATEMENT:Animal data suggest that reward-paired cues can promote maladaptive reward-seeking by biasing cost-benefit decision making. Whether this finding translates to humans is unknown. We examined the effects of salient reward-paired audio-visual cues on decision making under risk and uncertainty in human volunteers. Cues had risk-promoting effects on a risky choice task and independently increased task-related arousal as measured by pupil dilation. By demonstrating risk-promoting effects of cues in human participants, our data identify a mechanism whereby cue reactivity could translate into maladaptive behavioural outcomes in people with addictions.