A key prediction of ideomotor theories is that action perception relies on the same mechanisms as action planning. While this prediction has received support from studies investigating action perception in one-on-one situations, situations with multiple actors pose a challenge because in order to co-represent multiple observed actions, observers have to represent more actions in their motor system than they can physically execute. If representing multiple observed actions, like representing individual observed actions, recycles action planning processes, this should lead to response conflict by observation. In 5 experiments, we tested this hypothesis by investigating whether simply seeing two conflicting actions is sufficient to elicit response conflict and therefore adaptive control, in the same way as planning conflicting actions does. Experiments 1-3 provided meta-analytical evidence (N = 262) that seeing two conflicting gestures triggers a reverse congruency sequence effect on a subsequent, unrelated prime-probe task. Experiment 4 (N = 250) replicated this finding in a high-powered study. Finally, Experiment 5 (N = 253) revealed that the same effect was not present when using unfolding abstract shapes instead of moving hands. Together, these experiments show that not just planning but also seeing two conflicting actions elicits adaptive control and provide initial evidence that this is driven by motor conflict. These findings have important implications both for theories of action representation and research on cognitive control.