There have been numerous studies in which the biological role of oxytocin in trusting behavior has been investigated. However, a link between oxytocin and trust in humans was discovered only in one early study. We hypothesized that there is a large interindividual variation in oxytocin sensitivity, and that such variation is one reason for the doubt surrounding the role of oxytocin in trusting behavior. Here, in a double-blind, prospective, case-control study, we administered intranasal oxytocin to participants of trust and risk games. We measured salivary oxytocin concentration, relating it to the amount of money transferred among participants (a proxy for trust) and the autism-spectrum quotient (AQ). A one-sided Fisher’s exact test was performed to detect differences between the oxytocin and placebo groups in the proportions of investors who transferred the maximum amount of money. We discovered a tendency for participants who received oxytocin to transfer higher amounts of money to co-participants than those who received a placebo (P = 0.04). We also revealed a high degree of interindividual variation in salivary oxytocin concentrations after oxytocin administration. After stratifying the samples with respect to oxytocin sensitivity, oxytocin-sensitive participants in the oxytocin group also transferred higher amounts of money than those in the placebo group (P = 0.03), while such a tendency was not observed for oxytocin-insensitive participants (P = 0.34). Participants with lower AQ scores (less severe autistic traits) exhibited a greater tendency toward trusting behavior after oxytocin administration than did those with higher AQ scores (P = 0.02). A two-sample t-test that was performed to detect significant differences in the mean transfers between the oxytocin and placebo groups indicated no significant between-group difference in the mean transfers (P = 0.08). There are two possible interpretations of these results: First, there is no effect of oxytocin on trust in humans; second, the effects of oxytocin on trust in humans is person-dependent. However, the results should be interpreted with caution as the effect size was not larger than the minimal detectable effect size and the results were not statistically significant (P > 0.05) after Bonferroni corrections.