This paper investigates the effect of poverty and good intentions on dictator game giving. Previous experimental studies in which information was supplied to dictators about recipients have shown that dictator giving increases overall in this context. We develop a new design of standard informed dictator games with three main variants: 1) three recipients are used instead of one; 2) dictators are informed that their recipients are poor; 3) dictators give donations in the form of medicines instead of money. We have found that 46% of the experimental subjects (dictators) give the full amount of money (100% of the endowment) in the ‘poverty’ treatment, while in the ‘medicines’ treatment this percentage increases to 72%. Such extremely generous behavior has seldom been observed in the previous literature on dictator games.