Recent studies have reported on the utility of audio computer-assisted self-interviewing (ACASI) in surveys of human immunodeficiency virus (HIV) risk behaviors that involve a single assessment. This paper reports the results of a test of ACASI within a longitudinal study of HIV risk behavior and infection. Study participants (gay men (n = 1,974) and injection drug users (n = 903)) were randomly assigned to either ACASI or interviewer-administered assessment at their second follow-up visit 12 months after baseline. Significantly more of the sexually active gay men assessed via ACASI reported having sexual partners who were HIV antibody positive (odds ratio = 1.36, 95% confidence interval: 1.08, 1.72), and a higher proportion reported unprotected receptive anal intercourse. Among injection drug users (IDUs), our hypothesis was partially supported. Significantly more IDUs assessed via ACASI reported using a needle after another person without cleaning it (odds ratio = 2.40, 95% confidence interval: 1.34, 4.30). ACASI-assessed IDUs reported similar rates of needle sharing and needle exchange use but a lower frequency of injection. Participants reported few problems using ACASI, and it was well accepted among members of both risk groups. Sixty percent of the participants felt that the ACASI elicited more honest responses than did interviewer-administered questionnaires. Together, these data are consistent with prior research findings and suggest that ACASI can enhance the quality of behavioral assessment and provide an acceptable method for collecting self-reports of HIV risk behavior in longitudinal studies and clinical trials of prevention interventions.