Background and aims: Evidence for efficacy of brief alcohol interventions (BAIs) is mainly limited to primary care and at-risk drinkers. The aim was to test the efficacy of a BAI addressing the full spectrum of alcohol use in a general population sample and across alcohol risk groups. Design: Two-parallel-group randomized controlled trial (allocation ratio 1:1) with post-baseline assessments at months 3, 6 and 12. Setting: One municipal registry office in Germany responsible for registration, passport and vehicle admission issues. Participants: A total of 1646 proactively recruited 18–64-year-old adults with past year alcohol use (56% women, 66% low-risk drinkers) were randomized to intervention (n = 815) or control (n = 831). Intervention and comparator: The intervention consisted of assessment plus computer-generated individualized feedback letters at baseline and months 3 and 6. Comparator was assessment only. Measurements: Primary outcome was change in the self-reported number of drinks/week from baseline to 12 months. Changes at 3 and 6 months were secondary outcomes. Moderator was alcohol risk group (low-risk versus at-risk drinking) according to the Alcohol Use Disorders Identification Test–Consumption, with scores from 1-3 (women) and from 1-4 (men) indicating low-risk drinking. Findings: For the whole sample, significant group differences were observed neither at 12-month follow-up [incidence rate ratio (IRR) = 1.01, 95% confidence interval (CI) = 0.87–1.17, Bayes factor (BE) = 0.52] nor at previous assessments (month 3: IRR = 1.01, 95% CI = 0.92–1.12, BE = 0.41; month 6: IRR = 0.93, 95% CI = 0.81–1.07, BE = 1.10). Moderator analyses revealed that low-risk drinkers were more likely to benefit from BAI only at month 6 than at-risk drinkers (IRR = 0.77, 95% CI = 0.70–0.86). Conclusions: In a randomized controlled trial, there was no clear evidence for efficacy of a computer-based brief alcohol intervention in a general population sample, but there was some evidence of medium-term benefits in the large but understudied group of low-risk drinkers.