Once again, paternalism is rearing its ugly head. Whether in academia ? by behavioural economists of ?soft paternalism? conviction and more hardcore paternalist philosophers such as Sarah Conly ? or in politics ? by all those who advocate bans on meat, sugar, combustion engines, flying or, indeed, at times, leaving the house ? paternalist proposals have gained much popularity in recent years. In this paper I argue that an agent A?s interference with another agent P?s liberty is justified at best when three necessary (but not sufficient) conditions are met: 1. A knows what constitutes P?s well-being or what his goals or interests are. 2. A knows that Z promotes P?s well-being, goals, or interests. 3. Of all available actions that might promote P?s well-being, goals, or interests, Z is the action that least interferes with P?s liberty or autonomy. I argue that for most agents (and governments in particular) deciding about the implementation of possibly welfare-enhancing interventions, typically, conditions 1)-3) are not met