Liver transplantation (LT) for alcohol-associated hepatitis (AH) remains controversial due to concerns about candidate selection subjectivity, post-LT alcohol relapse and the potential exacerbation of LT disparities. Our aim was to design, perform and examine the results of a simulated selection of candidates for LT for AH. Medical histories, psychosocial profiles and scores, and outcomes of four simulation candidates were presented and discussed at two multidisciplinary societal conferences with real-time polling of participant responses. Candidate psychosocial profiles represented a wide spectrum of alcohol relapse risk. The predictive accuracy of four psychosocial scores, Dallas consensus criteria, SALT, SIPAT and QuickTrans, were assessed. Overall, 68 providers, mostly academic transplant hepatologists, participated in the simulation. Using a democratic process of selection, a significant majority from both simulations voted to accept the lowest psychosocial risk candidate for LT (72% and 85%) and decline the highest risk candidate (78% and 90%). For the two borderline risk candidates, a narrower majority voted to decline (56% and 65%; 64% and 82%). Two out of four patients had post-LT relapse. Predictive accuracies of Dallas, SIPAT and Quicktrans scores were 50%, while SALT was 25%. Majority voting outcomes were concordant with post-LT relapse in three out of four patients. When defining “success” in LT for AH, providers prioritized allograft health and quality of life rather than strict abstinence. In this simulation of LT for AH using a democratic process of selection, we demonstrate its potential as a learning model to evaluate the accuracy of psychosocial scores in predicting post-LT relapse and the concordance of majority voting with post-LT outcomes. Provider definitions of “success” in LT for AH have shifted towards patient-centered outcomes.