Article Preview
TopIntroduction
Cyber security is currently a high-ranking national security issue – a statement supported by recent congressional testimony noting that the United States saw a 782% increase in the number of reported cyber attacks against federal agencies from 2006 to 2012 (GAO-13-462T). Regarding potential causes of security breaches, the Ponemon Institute (Ponemon Institute 2013) suggested that 64 percent of data breaches in 2012 were the result of human error and problems in the ways that systems were constructed by humans (improperly configuring software that resulted in inadvertent data dumps, logic errors in data transfer, etc.). In a recent report by IBM that looked at common cyber attacks across 3,700 IBM clients in 130 countries, it was found that in most cases humans were the primary reason the breach occurred and humans were labeled as the ‘weak links’ in cyber networks (IBM 2013). The report also noted that cyber threats are becoming more opportunistic as human fallibility is exploited (IBM 2013), and the analysis suggested that human errors account for approximately 80 percent of company breaches.
With the exception of research studies devoted to cyber security training in specific settings (Abawajy 2012, Camp 2009, Jansson and von Solms 2013, Sheng et al. 2007), to the authors’ knowledge there is little empirical work exploring, articulating, or measuring the role of human operators during cyber breaches. The need for such empirical work has been the topic of several recent cyber research articles (Bowen et al. 2012, Boyce et al. 2011, Knott et al. 2013). More specifically, there is a need to understand the human-level trait and state factors at play when cyber attacks occur. To address this gap, the current paper views cyber attacks through the lens of suspicion. In order to reduce the human errors described above, computer users must learn to properly transition from normal working behavior to behavior under cyber attack (e.g., call IT, run antivirus software) at appropriate times. We hypothesize that suspicion plays an integral role as the conduit between these normal working behaviors and behaviors associated with detecting and appropriately reacting to a cyber attack.
This paper makes several contributions to the cyber security domain. We (i) describe and explore how the construct of suspicion operates during cyber attacks, (ii) empirically develop a suspicion-based, latent structure of cues that occur during cyber attacks, (iii) demonstrate how the derived latent structure can be used to develop and test hypotheses about the effects of those cues on users’ cognitive and emotional reactions, (iv) suggest and describe techniques to better train operators to detect, report, and appropriately react to security breaches, and (v) describe recent research with non-invasive physiological sensors that has the potential to monitor the mental states of operators in order to ensure optimum situation awareness in the cyber domain.