This paper argues that specific individual informed consent and other forms of consent predicated on a right to autonomy may not in all circumstances be appropriate for the establishment and use of large data sets of health information. We suggest that there are inherent failings in such an approach, shortcomings that we analyse below. We argue that individuals share an obligation to contribute their data, as doing so is cost-free (however cost is construed) and benefits accrue to the population as a whole. Large health data sets can be considered public goods -- goods that are non-rival in consumption (they are not depleted in being used, so others are not prevented from using them) and in some cases non-exclusive in use (e.g. when large datasets of non-identifiable information are publicly-accessible without cost, such as some Australian Bureau of Statistics data) -- and contributing to these goods may be an obligation, the meeting of which allows citizens to invest in knowledge infrastructure. The approach argued for here is a variation of a communitarian ethic in which people have an obligation to contribute their data but have no correlative right to expect or receive an individual benefit.