A 4-year randomized controlled trial found that computer-generated feedback letters failed to reduce alcohol consumption in the general population. Unexpectedly, the control group showed greater reduction at 48 months, suggesting that current digital intervention strategies may require significant adaptation for low-risk drinkers.