Artificial Phishes Bite Best: The Role of Agreeableness in Human- vs. Generative AI- generated Spear Phishing
Keywords
No Thumbnail Available
Authors
Issue Date
2025-07-01
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Generative AI models are increasingly used for phishing, as it makes attacks easier to conduct and personalize content better. Given that personality traits influence phishing susceptibility, this study investigates how the level of agreeableness influences individuals’ susceptibility to spear phishing emails by comparing their detection accuracy towards human- and generative AI-generated spear phishing emails. The primary objective was to assess whether agreeableness moderates detection accuracy across these email types. A between-subjects experiment was designed in which participants were asked to perform an email classification task for a persona they needed to emphasize with, to hide the true nature of the research. Participants were randomly classified to two groups: one viewed 14 mails of which 7 were generative AI-generated spear phishing emails, the other group saw 14 mails of which 7 were human-generated examples. Afterwards, participants completed personality assessments to investigate their level of agreeableness. The findings reveal that while emails made with generative AI were harder to detect, the level of agreeableness does not significantly influence detection accuracy between the two email types. The findings contribute to the cybersecurity field by showing the insignificant moderating role of agreeableness and the urgent need for improved defence mechanisms against rapid developing phishing attacks with generative AI.
Description
Citation
Supervisor
Faculty
Faculteit der Managementwetenschappen