Synthetic Media in Cybercrime: The Role of AI in Blackmail and Extortion

The rapid advancement of AI especially generative models such as Generative Adversarial Networks (GANs), Stable Diffusion Models, Large Language Models and Large Visual Models has led to significant breakthroughs in the creation of highly realistic synthetic media. While these technologies offer several beneficial applications, they also introduce new opportunities for malicious activities, including AI powered blackmailing and extortion attacks. This thesis explores the escalating threat of generative blackmailing, where advanced generative artificial intelligence technologies are exploited to create realistic synthetic media aimed for extortion or reputation harm. The study begins with a thorough examination of these generative models, followed by an analysis of deepfake technology and its role in facilitating new forms of blackmail. After conducting an experiment to explore the mechanisms of producing synthetic media and creating a sample using deepfake technology, the research develops a systematic categorization of specific case studies. This research identifies several types of generative blackmailing attack instances: corporative/business threat, individual/reputation, and political/mis-informative. An extensive impact analysis evaluates the psychological, social, legal, and financial consequences of these attacks on individuals, businesses, and public opinion. Moreover, the thesis assesses the effectiveness of current detection techniques and proposes advanced preventive strategies to counteract these emerging threats, with recommendations provided to guide the development of regulatory frameworks that can mitigate the risks associated with generative technologies. The findings of this research contribute significantly to the field of cybersecurity by offering a deeper understanding about the mechanisms of AI-driven threats and presenting practical measures to safeguard against the nefarious use of synthetic media. Additionally, presenting a formal framework of categorizing these attacks contributes to the legal, social and technological landscapes as it can serve as a foundation for drafting modern legislation that aligns with emerging threats and lays the groundwork for advancing detection and prevention methodologies in the future.