Exploitation is commonly understood as taking unfair advantage. This article discusses the various prominent accounts that have been offered of how an exchange, despite being Pareto improving and consensual, can nevertheless count as unfair or unjust and, hence, as presumptively impermissible. Does the wrongness of an exploitative transaction consist in its compounding a prior distributive injustice, or in its deliberately profiting from someone’s vulnerability, or in its commodification of that which should not be commodified? How should responsibility for exploitation be assigned, and can this avoid generating moral hazard? The accounts of exploitation analysed here are classified along two dimensions—historical vs. ahistorical and intentional vs. non-intentional—in their conceptions of unfairness, and the possibility of a hybrid account is explored.