Some argue that human groups have a stake in the outcome of population-genomics research and that the decision to participate in such research should therefore be subject to group permission. It is not possible, however, to obtain prior group permission, because the actual human groups under study, human demes, are unidentifiable before research begins. Moreover, they lack moral standing. If identifiable social groups with moral standing are used as proxies for demes, group approval could be sought, but at the expense of unfairly exposing these surrogates to risks from which prior group approval is powerless to protect them. Unless population genomics can proceed without targeting socially defined groups, or can find other ways of protecting them, it may fall to individuals to protect the interests of the groups they care about, and to scientists to warn their subjects of the need to do so.