Examining generational impacts of machine learning, artificial intelligence, generative AI and deepfake-based social engineering cybercrime

Abstract

Social engineering (SE) has been used as an effective technique for decades to extract information from individuals for malicious purposes. Machine learning (ML), artificial intelligence (AI), generative AI (GenAI) and deepfakes provide advanced technologies and tools for cybercriminals to increase the probability of a successful cyberattack. Our ability to detect that we are being maliciously manipulated is becoming increasingly difficult as these technologies are incorporated into SE-based cyberattacks. This paper examines the positive and negative aspects of ML, AI and GenAI, reviews existing literature about these technologies and presents a brief introduction of the technical aspects of ML, AI and deepfakes. The decision model Required Elements for a Social Engineered Cyber Attack Theory examines our decision processes when confronted by a SE-based cyberattack, and how simple decisions can become more complicated when influenced by AI. Examples of generation-specific AI-based scams are examined to describe techniques bad actors are using to victimise specific age groups. The conclusion offers recommendations that technology users can adopt to reduce their exposure of being victimised by AI-based SE cyberattacks.