<p>With the rapid technological advancements, Deepfakes have emerged as a powerful cyber-threat utilizing deep learning algorithms to generate highly realistic synthetic media that depict events or situations that never actually occurred. The widespread use of social media platforms and the increasing availability of individuals’ biometric data online have made large quantities of training data easily accessible, making it possible for everyone to become the subject of a Deepfake. Despite its potential for various beneficial applications, the use of Deepfakes without consent or knowledge of depicted individuals has created new avenues for the commission of numerous criminal acts. This study provides a comprehensive analysis of Deepfake technology, evaluating its evolution from foundational algorithms to its emergence as a new cybersecurity threat. Adopting a threat-centric framework, the paper distinguishes between Deepfakes as cyber-enabled crimes and cyber-dependent crimes. Furthermore, the study also discusses the defence asymmetry in terms of ongoing regulatory frameworks, technical capabilities and awareness factors to facilitate an understanding on how Deepfakes pose a growing threat and offers recommendations for tackling effects of malicious Deepfakes.</p>