ABSTRACT
The role of technology has undoubtedly evolved into amplifying attackers' ability to use the cyber space for the deceit and abuse of Internet users. This dissertation seeks to investigate these problems from the lens of deceptive and abusive content (e.g., phishing, social engineering, dis- or misinformation, intimate partner surveillance). Phishing is an extremely popular cyber-social engineering attack that come with great costs to society-at-large, and along with mis- and disinformation, has risen to society's collective consciousness after the 2016 and 2020 U.S. General Elections, as well as the onset of the COVID-19 pandemic. Furthermore, research into crimes of abuse surrounding intimate partner violence (IPV), colloquially known as domestic abuse, is still in its infancy, yet IPS allows abusers to stalk, monitor, intimidate, and harass their victims as a form of further control. This thesis seeks to tackle these issues through a multi-methods approach, including natural language processing to detect the presence of influence cues in text, qualitative methods, and rigorous statistical analyses. I detail how cyber abuse is leveraged in social media, how expert advice can negative affect minorities, and investigate how subtle online toxicity can be automatically detected. I then investigate several different ways to mitigate the harms of online deception. The work detailed in this dissertation has resulted in novel and publicly available datasets that may pave the way for further novel mitigations and solutions within their fields and to the global problem of cyber deception and abuse. (PsycInfo Database Record (c) 2022 APA, all rights reserved)
ABSTRACT
The role of technology has undoubtedly evolved into amplifying attackers' ability to use the cyber space for the deceit and abuse of Internet users. This dissertation seeks to investigate these problems from the lens of deceptive and abusive content (e.g., phishing, social engineering, dis- or misinformation, intimate partner surveillance). Phishing is an extremely popular cyber-social engineering attack that come with great costs to society-at-large, and along with mis- and disinformation, has risen to society's collective consciousness after the 2016 and 2020 U.S. General Elections, as well as the onset of the COVID-19 pandemic. Furthermore, research into crimes of abuse surrounding intimate partner violence (IPV), colloquially known as domestic abuse, is still in its infancy, yet IPS allows abusers to stalk, monitor, intimidate, and harass their victims as a form of further control. This thesis seeks to tackle these issues through a multi-methods approach, including natural language processing to detect the presence of influence cues in text, qualitative methods, and rigorous statistical analyses. I detail how cyber abuse is leveraged in social media, how expert advice can negative affect minorities, and investigate how subtle online toxicity can be automatically detected. I then investigate several different ways to mitigate the harms of online deception. The work detailed in this dissertation has resulted in novel and publicly available datasets that may pave the way for further novel mitigations and solutions within their fields and to the global problem of cyber deception and abuse. (PsycInfo Database Record (c) 2022 APA, all rights reserved)