Year: 2024
Type: dissertation
Abstract: This thesis explores the impact of generative AI on enhancing cybercriminal activities, particularly through the creation of fraudulent content that affects victim perception. A series of formal experiments (n = 15) assessed how AI-generated fraudulent materials compare to human-made frauds in influencing victims, using the elaboration likelihood model (ELM) and related cognitive research. The experimental design involved within-group comparisons, focusing on response time, authenticity, credibility, as well as individual background and self-assessment. Findings suggest that AI-generated content, while typically perceived as less credible than authentic materials, matches the persuasiveness of human-made frauds under certain conditions. This highlights the concerning potential of generative AI in cybercrime, capable of emulating human deceptive tactics effectively. The thesis highlights the necessity for innovative cybersecurity strategies and an informed society to address the challenges posed by AI-enhanced cyber fraud. It advocates for continued research and updated regulatory measures to mitigate the evolving threats of AI in cybercriminal exploit.
Authors: Bjelvér, Jakob, Welsapar, Michael
Institution: Lund University
Language: English
Subfield: Law
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7029336371