The study explores the relationship between digital literacy, exposure to AI-generated deepfake videos, and the ability to identify deepfakes by Generation X in Indonesia who are currently between the ages of 43 and 58. It also analyzes the impact of deepfake identification capabilities on the cognitive, affective, and behavioral aspects of internet users. Through a survey involving 199 respondents taken from a total population of 42 million Generation X internet users in Indonesia, it applied a random sampling method. The sample size was determined by the Slovin formula with a confidence level of 90% and a margin of error of 7.1%. The descriptive analysis shows a moderate level of digital literacy and relatively low exposure to deepfakes. However, the ability to identify deepfakes was found to be low. The results of inferential statistical analysis show that digital literacy and exposure to deepfakes do not have a significant influence on the ability to identify deepfakes. Additionally, the ability to identify deepfakes does not significantly affect cognition, compassion, or behavior. While digital literacy is important, these findings reinforce the assumptions of Generation Theory and Media Dependency Theory. Additionally, it suggests that specific training on media manipulation technologies is needed to improve deepfake detection capabilities. This research implies that efforts to improve digital literacy should be expanded, including technical skills and critical thinking relevant to manipulative media such as deepfakes.