Phishing Email Detection Through Amygdala Hijack Threats

Phishing attacks, a social engineering method to gain sensitive information from a victim, have grown in popularity. Previous literature has shown that phishers use various psychological methods in order to influence a victim. However, the effectiveness of threat cues and its emotional effect have rarely been investigated. The theoretical premise of this paper is that threat cues stimulate the amygdala region of the brain and influence phishing email recipients into clicking behaviors. We focus on amygdala hijack threat cues of punishment threat, territory threat, and status threat, and address how effective amygdala hijack threats are in detecting phishing emails. In this paper, we created a machine learning model to classify phishing emails based on the presence of threat cues and compared it to a baseline model without threat cues. We observed that in phishing email detection, models with the presence of threat cues significantly outperformed the baseline models by around 3% to 8% in terms of F-score. A practical implication of this study is to create an anti-phishing solution to classify phishing emails based on amygdala hijack threats and their emotional impact on recipients of phishing emails, allowing for innovations in effective countermeasures against phishing