In two experiments we investigated whether older adult controls (OACs) and people with mild and moderate Alzheimer’s disease (AD) benefit from false memory priming effects in subsequent problem-solving tasks. In addition, and unlike in previous false memory priming studies with older adults, we examined latency measures in the recognition phase. In Experiment 1 participants were asked to solve compound remote associate task (CRAT) problems, half of which had been preceded by the presentation of Deese/Roediger-McDermott (DRM) lists whose critical lures (CLs) were also the solutions to those problems. In Experiment 2, we used a similar paradigm but investigated whether CLs could prime solutions to subsequent analogical reasoning problems. In this latter experiment, we also examined whether these priming effects were stronger when the activation of the CL term occurred during the memory task (was presented as part of the list; i.e., true memories) or when these items were not presented but arose during encoding due to spreading activation (i.e., false memories). We found that all three groups’ performance on these tasks was facilitated only by false memories spontaneously generated from the prior presentation of DRM lists. That is, performance on CRATs and analogical reasoning tasks was better (greater accuracy and faster speed) when those problems were preceded by DRM lists whose CLs also served as the solution to those problems. These findings are consistent with previous results from studies with children, young adults, and older adults and extends them to people with more moderate AD.Copyright © 2020 Elsevier Ltd. All rights reserved.