It is currently debated whether simple forms of social perspective-taking that are in place by late infancy are performed automatically. We conducted two experiments (
N
= 124) to test whether 3-year-olds show automatic perspective-taking during object searches, and whether automatic perspective-taking is facilitated by joint attention. Children were asked to retrieve an object immediately after it was moved from one (L1) to another (L2) location within a container, e.g. a sandbox. In Experiment 1, a between-subjects design was used, with children being randomly assigned to one of three experimental conditions: one in which child and other jointly attended to the object in L1 (joint attention condition); one in which the other was present but unengaged with the child when the object was placed in L1 (other present condition) and a baseline condition in which only the child was present (no other condition). Automatic perspective-taking should manifest in biased searches toward L1 in the other present and joint attention conditions, but not in the no other condition. No automatic perspective-taking was observed in either experiment, regardless of whether the other person left and remained absent (Experiment 1) or returned after the object was relocated (Experiment 2). The findings contribute to a growing body of empirical data that questions the existence of automatic perspective-taking.