Fuzzy-trace theory (FTT) posits that there are two independent types of memory processing: gist and verbatim. Meaning-based, gist memory, unlike detailed, verbatim memory, is more resistant to fading over time, but this resilient type of memory can also facilitate the production of false memories. This is especially true when the items to be remembered are connected via semantic relationships, as in stereotypes, and reliance on gist memories, increases with age. We first provide an overview of FTT and the false memory phenomenon. We then present results from 2 experiments using a cued recall task tapping self-generated stereotypes for eating habits of two hypothetical people (a healthy eater and an unhealthy eater) based on the frequency of their meals. In the first experiment, we used a between-subjects manipulation of delay (either immediately tested or after 7 days) and recruited 2 age groups (18-22 or older than 55). 770 college-age and 92 older adults in the U.S. were shown 20 meals that two hypothetical people ate in the past month. We then tested participants on how many times a certain meal repeated among the 20 they studied. Probes included targets, gist-consistent distractors and gist-inconsistent distractors. This design allowed us to gather data on how false memories for self-generated stereotypes compared to true memories for each age group and how that changed with or without a delay. In our second study, we again examined the effects of delay and age with a sample of 117 college-aged (18-22) and 133 post college (23-55) participants from Brazil. Last, we ran a planned comparison of the 770 participants in the 18-22 group from the U.S. and the 133 participants from the 18-22 group from Brazil for another 2 (country) x 2 (immediate or delay) between-subjects design. In our experiments, results supported FTT’s predictions: Related distractors that fit the stereotype were falsely remembered to have appeared as often as targets that were shown. This effect grew stronger with age and after a delay, such that our oldest age group, over 55 years of age, estimated stereotype-consistent distractors as having been presented multiple times and, in some cases, more often than the true targets. Our results show that self-generated stereotypes not only facilitate false memory production, but that this effect is reliable across cultures and is strengthened as one ages and with a delay between presentation and testing. These results are particularly important in a legal context, where memory is rarely tested immediately after important events and accurate memory can determine if a defendant is convicted.