Trustworthiness plays a crucial role in facilitating effective and successful interactions. For example, perceived trustworthiness influences our behavior when investing (van ’t Wout & Sanfey, 2008) and during strategic decision-making (Johnson & Mislin, 2011). Interestingly, perceived trustworthiness plays a critical role in initial interactions, but its influence becomes smaller over time (Chang et al., 2010). Here, we investigated the impact of perceived trustworthiness on a crucial aspect of the human adaptive behavior, value-based decision-making. In two experiments, participants played a modified version of "The Door Game," in which they searched for a reward behind one of two doors based on the advisors' advice (Van der Biest et al., 2020). The advisors were computer-generated faces classified as either trustworthy or untrustworthy (Oosterhof and Todorov, 2008). Crucially, the advisor’s validity was the same for both levels of trustworthiness (i.e., 66%). In Experiment 1 (N=150), we found a higher rate of advice following for trustworthy advisors, but only during the first half of the experiment. In our preregistered replication (Experiment 2, N = 300), we found a similar pattern for the advice following rate, thereby we successfully replicated the results obtained in Experiment 1. In addition, we found that during the initial encounters (i.e., block one) participants were faster to decide when the advisor was trustworthy compared to untrustworthy. Overall, our results highlight the reliance on social information in value-based decision-making, while additionally, emphasizing rapid adjustments of trust beliefs and its dynamic nature over time.