Accounts of consent in medical ethics typically assume that consent plays the same role irrespective of the type of treatment. In this paper I argue that this assumption is false. Because of this, obligations to provide information to patients that stem from the need for consent to be valid will not apply to all types of treatment. This does not mean that there are no reasons to provide such information. The second part of the paper maps out what these reasons are and argues that they are grounded in the obligation of beneficence and a duty to warn, not in considerations of respect for autonomy.