Event-related potentials (ERPs) and performance measures were recorded from younger and older adults during two working memory tasks. In both tasks, digits were presented in strings that varied in length from 4-11 numbers. In one task, participants were instructed to remember the last four digits (Updating Task, UT) and in the other task the first four digits (delayed-matching-to-sample task, DMST) for a subsequent memory test. For both age groups, performance for the DMST was better than for the UT, but younger adults performed better than the older adults on the UT. Three major findings were observed in the ERP data. First, under DMST conditions, ERP indices of inhibition for task-irrelevant stimuli were greater for younger than older adults, even though performance did not differ between the age groups. Second, under UT conditions, ERP indices of pro-active interference were observed for younger but not older adults. Therefore, when task demands were constrained, older adults were still able to maintain performance even in the face of reduced inhibition of task-irrelevant stimuli. Third, for both age groups, a positive component (peak latency ~500 ms) of maximum amplitude was elicited by the 5th study digit under UT conditions, and the 4th digit under DMST conditions. Current source density maps showed a stronger frontal focus for the older than the younger adults under the UT conditions. These findings suggest that the positive ERP component reflects processing of salient task events rather than updating per se, and that older adults rely more on frontal lobe processes.