Even from early childhood, humans have a sense of fairness. Unfair players are judged harshly and people expect those who shirk effort at the expense of others to be punished, not rewarded. As adults, there is also a reputational motivation to be seen as fair, and people prefer to continue interactions with fair interaction partners rather than unfair ones. However, to date these studies have examined fairness in a third-party perspective: they often look at how people judge other people’s fairness in relation either to themselves or to other people. These studies have also focussed on resource or reward distribution in response to fair or unfair behaviour. There is little research looking at how and whether people choose to be fair themselves when they are responsible for task distribution. It is also unclear whether a desire to be fair could compete with other motivations in joint tasks, namely efficiency. Previous research has shown that interactants show a strong drive towards maximising the overall efficiency of joint actions (co-efficiency), even when this means sacrificing the efficiency of their individual actions. In the present study, participants completed a task where they were passing objects on a screen to a virtual partner whom they believed to be a real person in a different room. Participants could choose to pass the object at one of two points – one that was equidistant between the origin and target site, meaning that both actors would contribute the same effort in moving the item (fair) and one where the passing site was closer to either the origin or target (unfair). The unfair route could be either longer overall than the fair route (fairness and co-efficiency are congruent), shorter overall (fairness and co-efficiency are incongruent), or the same length (equal trials). We find that people show a strong drive towards maximising co-efficiency, choosing the shortest overall path even when this disadvantages themselves, but little evidence for a drive towards fairness. On equal trials, participants show evidence of altruism in their decisions, often choosing the route that minimises the contribution of their partner when this does not compromise the overall co-efficiency. As such, it is not that participants behave selfishly during interactions, but that they do not have a drive to ensure a fair task distribution on a trial-by-trial basis. Instead, they choose to be efficient as a group.