We tested the hypothesis that there is an age-related deficit in the recovery from interruptions. This hypothesis is based on the idea that it is more difficult for old than for young adults to establish a focused state after working memory has been “opened up” through an interruption. Old (N = 95) and young adults (N = 94) performed competing nondominant and dominant primary tasks (selecting either exogenously or endogenously cued targets) in alternating, single-task blocks that were occasionally interrupted through trials with unrelated math tasks. As predicted, after interruptions, older adults showed increased and prolonged recovery costs, as well as generally larger endogenous/exogenous and conflict effects in blocks that contained interruptions. Individual differences in working memory (WM) capacity did not produce comparable results, suggesting that the interruption-based deficits were specific to aging. In addition, the theoretically important, paradoxical cost asymmetry (i.e., larger interruption costs for the dominant exogenous than for the nondominant endogenous task) was stronger in old than in young adults. These results provide novel insights about the interplay between WM and long-term memory during task control, as well as the origin of age differences in task-set selection. (PsycInfo Database Record (c) 2022 APA, all rights reserved)