We conducted a longitudinal study of participants in phase I and II HIV vaccine safety and immunogenicity trials to examine changes in sexual risk behavior that are associated with risk of HIV transmission. The participants were 48 HIV-negative men and women enrolled in one of two placebo-controlled HIV vaccine trials conducted at San Francisco General Hospital. There was a significant increase in insertive unprotected anal intercourse (UAI) from 9% at baseline (trial entry), to 13% at the month 6 assessment, to 20% at the month 12 assessment (p = .02). The primary predictor of either insertive or receptive UAI during the vaccine trials was having engaged in this behavior prior to entry (p = .001). Higher-risk behavior was also seen among participants who were younger and had multiple sexual partners (each, p = .06) and who indicated that one of their reasons for participation in the vaccine trial was hope of protection from HIV infection (p = .07). These findings indicate that despite instructions otherwise, participants with a history of high-risk behavior or who express hope of protection from HIV infection by enrolling in vaccine trials may be candidates for more intensive risk-behavior counseling prior to and during their participation.