This review focuses on 27 studies employing experimental alcohol self-administration (ASA) in humans which were published between 1989 and 2010. Twelve studies enrolling healthy, non-dependent social drinkers (HSD) were aimed at evaluating physiological and behavioral determinants of alcohol-induced reward or modeling situations of increased risk to develop alcohol use disorders. The remaining 15 studies tested the effect of medications such as naltrexone, nalmefene, nicotine, mecamylamine, varenicline, gabapentin, aripiprazole, and rimonabant on ASA. The participants were either HSD or non-treatment-seeking alcoholics (NTSA). In 25 of these studies, the subjects ingested alcohol orally and reached a mean peak blood alcohol concentration (BAC) during baseline conditions between 43 and 47 mg% (0.043-0.047%). Two recent studies employed computer-assisted self-infusion of ethanol (CASE), where subjects press a button to request multiple sequential alcohol exposures intravenously instead of drinking. This method has been demonstrated to be safe and provides increased experimental control of BAC and keeps subjects blind concerning the amount already self-administered. Peak exposures in the CASE studies ranged from 60 to 80 mg% in HSD and up to 240 mg% in NTSA.