University of Twente Student Theses
Leveraging Generative Pre-trained Transformers for the Detection and Generation of Social Engineering Attacks : A Case Study on YouTube Collusion Scams
Perik, L.W. (2025) Leveraging Generative Pre-trained Transformers for the Detection and Generation of Social Engineering Attacks : A Case Study on YouTube Collusion Scams.
|
PDF
2MB |
| Abstract: | Social engineering remains a significant threat in the digital age, with recent statistics by PurpleSec revealing that over 98% of cyberattacks involve some kind of social engineering. At the same time, a recent report by Europol reveals the trend of Large Language Models (LLMs) increasingly being exploited to facilitate these attacks. Consequently, this study examines the dual applicability of pre-trained LLMs in the social engineering landscape, focusing on both the potential of LLMs to be abused by adversaries to generate social engineering attacks and their utility in enhancing the detection and mitigation of such schemes. Our findings firstly demonstrate that pre-trained LLMs like GPT-4 can easily be exploited to generate sophisticated collusion scams and bypass existing detection mechanisms. Notably, GPT-4 consistently complied with our requests, highlighting the potential for misuse by adversaries and the need for more robust countermeasures. Additionally, we find that LLMs can improve the detection and mitigation of these social engineering attacks, with our methodology demonstrating versatility in accurately being able to detect both LLM-generated and real collusion scams. Lastly, we conclude the paper with a call for continued research into the application of LLMs in this regard, highlighting multiple valuable areas for improvement and future work. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science MSc (60300) |
| Link to this item: | https://purl.utwente.nl/essays/104867 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page