@deli_rum selection bias. Suppose the opposite were true: it's usually a greater personal risk to do the wrong thing. In the majority of cases then, people who do a thing because it's right and people who do a thing because the incentive structure rewards it have no disagreement. There's little point in a deep discussion of "the right thing" in circumstances where the alternative is obviously stupid. So the expression would be disproportionately used in that minority of situations where that's not the case, and if you only look at those situations you'll see the same trend as if the personally risky thing was in fact usually synonymous with right. So the observation therefore isn't strong evidence in either direction.

TL;DR: The conditions under which the expression is used aren't necessarily representative of all the conditions created by society's incentive structures.