While research indicates that benzodiazepine (BZD)-like drugs impair driving performance, it remains unclear (i) how far BZDs affect lane-keeping performance, compared with alcohol and (ii) to what extent this impact can realistically be measured in a simulated environment. To clarify these issues, 16 healthy male drivers who had never previously taken BZDs underwent a randomized, crossover, double-blind, placebo-controlled driving paradigm (with the BZD lorazepam) in both real-world and simulated settings. Two lane-keeping variables, namely inappropriate line crossings (ILCs) and standard deviation of lateral position (SDLP), were recorded during the driving sessions. Analyses revealed that (i) a single lorazepam dose (2mg given by mouth) caused higher SDLP increases than a blood alcohol concentration of above 0.05%, and that (ii) this BZD effect was amplified in the simulated driving setting, mainly for ILCs. As a consequence, we recommend that physicians be made more aware of BZD-related risks and that researchers make a clear distinction between the effects of BZD intake per se and the impact of simulated driving settings.