By Isabella Voce and Anthony Morgan

This study analyses data from a survey of Australian adult computer users conducted in June 2021 to examine the influence of online routine activities and life stressors on the likelihood of profit-motivated cybercrime victimisation.

Compared with non-victims, victims spent more time online, more frequently engaged in recreational online activities and were more likely to employ higher-risk online practices. Small-to-medium enterprise owners working from home were more likely to be victims. Respondents who had experienced recent increases in financial stress and gambling and negative impacts on interpersonal relationships during the COVID-19 pandemic were also more likely to be a victim of cybercrime.

Being accessible online and a lack of personal and physical guardianship are associated with an increased risk of being a victim, but other factors may influence the susceptibility of computer users to cybercrime victimisation. This has important implications for cybercrime responses

Trends & issues in crime and criminal justice no. 675. Canberra: Australian Institute of Criminology. 2023. 18p.

By Felix Soldner, Bennett Kleinberg, Shane D. Johnson

The popularity of online shopping is steadily increasing. At the same time, fake product reviews are published widely and have the potential to affect consumer purchasing behavior. In response, previous work has developed automated methods utilizing natural language processing approaches to detect fake product reviews. However, studies vary considerably in how well they succeed in detecting deceptive reviews, and the reasons for such differences are unclear. A contributing factor may be the multitude of strategies used to collect data, introducing potential confounds which affect detection performance. Two possible confounds are data-origin (i.e., the dataset is composed of more than one source) and product ownership (i.e., reviews written by individuals who own or do not own the reviewed product). In the present study, we investigate the effect of both confounds for fake review detection. Using an experimental design, we manipulate data-origin, product ownership, review polarity, and veracity. Supervised learning analysis suggests that review veracity (60.26–69.87%) is somewhat detectable but reviews additionally confounded with product-ownership (66.19–74.17%), or with data-origin (84.44–86.94%) are easier to classify. Review veracity is most easily classified if confounded with product-ownership and data-origin combined (87.78–88.12%). These findings are moderated by review polarity. Overall, our findings suggest that detection accuracy may have been overestimated in previous studies, provide possible explanations as to why, and indicate how future studies might be designed to provide less biased estimates of detection accuracy. 

PLoS ONE 17(12): 2022

By Sergi D. Bray , Shane D. Johnson and Bennett Kleinberg

Deepfakes’ are computationally created entities that falsely represent reality. They can take image, video, and audio modalities, and pose a threat to many areas of systems and societies, comprising a topic of interest to various aspects of cybersecurity and cybersafety. In 2020, a workshop consulting AI experts from academia, policing, government, the private sector, and state security agencies ranked deepfakes as the most serious AI threat. These experts noted that since fake material can propagate through many uncontrolled routes, changes in citizen behaviour may be the only effective defence. This study aims to assess human ability to identify image deepfakes of human faces (these being uncurated output from the StyleGAN2 algorithm as trained on the FFHQ dataset) from a pool of non-deepfake images (these being random selection of images from the FFHQ dataset), and to assess the effectiveness of some simple interventions intended to improve detection accuracy. Using an online survey, participants (N = 280) were randomly allocated to one of four groups: a control group, and three assistance interventions. Each participant was shown a sequence of 20 images randomly selected from a pool of 50 deepfake images of human faces and 50 images of real human faces. Participants were asked whether each image was AI-generated or not, to report their confidence, and to describe the reasoning behind each response. Overall detection accuracy was only just above chance and none of the interventions significantly improved this. Of equal concern was the fact that participants’ confidence in their answers was high and unrelated to accuracy. Assessing the results on a per-image basis reveals that participants consistently found certain images easy to label correctly and certain images difficult, but reported similarly high confidence regardless of the image. Thus, although participant accuracy was 62% overall, this accuracy across images ranged quite evenly between 85 and 30%, with an accuracy of below 50% for one in every five images. We interpret the findings as suggesting that there is a need for an urgent call to action to address this threat. 

Journal of Cybersecurity, 2023, 1–18 

By Shane D. Johnson and  Manja Nikolovska

Objectives Routine activity theory suggests that levels of crime are affected by peoples’ activity patterns. Here, we examine if, through their impact on people’s on- and off-line activities, COVID-19 restriction affected fraud committed on- and off-line during the pandemic. Our expectation was that levels of online offending would closely follow changes to mobility and online activity—with crime increasing as restrictions were imposed (and online activity increased) and declining as they were relaxed. For doorstep fraud, which has a different opportunity structure, our expectation was that the reverse would be true. Method COVID-19 restrictions systematically disrupted people’s activity patterns, creating quasi-experimental conditions well-suited to testing the effects of “interventions” on crime. We exploit those conditions using ARIMA time series models and UK data for online shopping fraud, hacking, doorstep fraud, online sales, and mobility to test hypotheses. Doorstep fraud is modelled as a non-equivalent dependent variable, allowing us to test whether findings were selective and in line with theoretical expectations. Results After controlling for other factors, levels of crime committed online were positively associated with monthly variation in online activities and negatively associated with monthly variation in mobility. In contrast, and as expected, monthly variation in doorstep fraud was positively associated with changes in mobility. Conclusions We find evidence consistent with routine activity theory, suggesting that disruptions to people’s daily activity patterns afect levels of crime committed both on- and off-line. The theoretical implications of the findings, and the need to develop a better evidence base about what works to reduce online crime, are discussed. 

Journal of Quantitative Criminology, 2022.

By John Aloysius, Ankur Arora, Viswanath Venkatesh

Purpose: Retailers are implementing technology-enabled mobile checkout processes in their stores to improve service quality, decrease labor costs and gain operational efficiency. These new checkout processes have increased customer convenience primarily by providing them autonomy in sales transactions in that store employee interventions play a reduced role. However, this autonomy has the unintended consequence of altering the checks and balances inherent in a traditional employee-assisted checkout process. Retailers, already grappling with shoplifting, with an estimated annual cost of billions of dollars, fear that the problem may be exacerbated by mobile checkout and concomitant customer autonomy. The purpose of this paper is to understand the effect of mobile checkout processes in retail stores on cybercrime in the form of shoplifting enabled by a technology transformed the retail environment. Design/methodology/approach The authors conducted an online survey of a US sample recruited from a crowdsourced platform. The authors test a research model that aims to understand the factors that influence the intention to shoplift in three different mobile checkout settings − namely, smartphone checkout settings, store-provided mobile device checkout settings, and employee-assisted mobile checkout settings − and compare it with a traditional fixed location checkout setting. Findings: The authors found that, in a smartphone checkout setting, intention to shoplift was driven by experiential beliefs and peer influence, and experiential beliefs and peer influence had a stronger effect for prospective shoplifters when compared to experienced shoplifters; in a store-provided mobile devices checkout setting, experiential beliefs had a negative effect on shoplifters’ intention to shoplift and the effect was weaker for prospective shoplifters when compared to experienced shoplifters. The results also indicated that in an employee-assisted mobile checkout setting, intention to shoplift was driven by experiential beliefs and peer influence, and experiential beliefs had a stronger effect for prospective shoplifters when compared to experienced shoplifters. Originality/value: This study is the among the first, if not first, to examine shoplifters’ intention to shoplift in mobile checkout settings. We provide insights into how those who may not have considered shoplifting in less favorable criminogenic settings may change their behavior due to the autonomy provided by mobile checkout settings and also provide an understanding of the shoplifting intention for both prospective and experienced shoplifters in different mobile checkout settings.

Information Technology and People, April 2019, 32(5):1234-1261