Good observation. When X >=a, every random value of X is >=a. Therefore, the mean of X has to be >= a. The random variable I can take only two values: 0 or 1. Therefore, the random variable aI can take only two values: 0 or a, both of which are <= X. Therefore mean of aI is necessarily <= the mean of X. In other words, E(aI) <= E(X). Sorry if this wasn't brought out clearly enough.