International
law enforcement agencies report an apparent relationship between autism
and cyber-dependent crime, although any such link remains unproven
(Ledingham and Mills 2015; NCA 2017).
This was the first study to empirically explore whether autism,
autistic-like traits, explicit social cognition, interpersonal support
and digital skills were predictors of cyber-dependent criminality.
Whilst higher levels of autistic-like traits were associated with a
greater risk of committing cyber-dependent crime, a self-reported
diagnosis of autism was associated with a decreased risk of committing
cyber-dependent crime. Around 40% of the association between
autistic-like traits and cyber-dependent crime was attributable to
greater levels of advanced digital skills. Basic digital skills were
also found to be a mediator between autistic-like traits and
cyber-dependent crime, although they accounted for a smaller proportion
of the association than advanced digital skills.
These
findings are consistent with the proposal that the apparent association
between autism and cyber-dependent crime identified by law enforcement
agencies may be reflecting higher levels of autistic-like traits amongst
cybercriminals but that this does not necessarily equate to autism
being a risk factor for cybercrime. This confusion may well arise
because typically, autistic people do report higher levels of
autistic-like traits than the general population (Ruzich et al. 2015a).
Cyber-dependent crime may therefore represent an area that
distinguishes high autistic-trait non-autistic groups from autistic
groups, consistent with proposal that people with autism differ
qualitatively from non-autistic people who are nevertheless high in
autistic-like traits (see Ashwood et al. 2016; Frith 2014).
The finding that autistic respondents were less likely to commit
cyber-dependent crime is also consistent with literature suggesting that
autistic people are generally as law abiding, if not more so, than the
general population. Lower levels of criminality are shown, at least for
certain types of crime (Blackmore et al. 2017; Cheely et al. 2012; Ghaziuddin et al. 1991; Heeramun et al. 2017; Howlin 2007; King and Murphy 2013; Murrie et al. 2002; Wing 1981; Woodbury-Smith et al. 2005a, 2006; but see, Rava et al. 2017; Tint et al. 2017).
Thus,
there is evidence that higher AQ scores are associated with higher
levels of cyber-dependent crime regardless of an autism diagnosis. As
this association was independent from the autism diagnosis, there may be
something about autistic-like traits beyond the diagnostic criteria for
autism that relates to cyber-dependent criminal activity. The mediation
analysis suggests that an association between autistic-like traits and
advanced digital skills may represent a key factor. We cautiously state
above that those reporting an autism diagnosis were less likely to
report cyber-dependent crime. Cautiously, as this could be for various
reasons beyond high AQ and autism being different things, including a
diagnosis of autism leading to some protection (e.g., more support
leading to less potential criminal behaviour; see Heeramun et al. 2017).
Importantly, however, there are potential selection issues in relation
to individuals who respond to an invitation to complete an online survey
on this topic, thus the possibility of selection bias cannot be ruled
out. We do not know how many did not respond to the invitations (and
therefore could not identify a response rate, for example) and the
apparent protective effect could be a chance finding due to small
numbers. Future research using larger samples can address such concerns
and until that time the suggestion that autism may be protective should
be considered speculative, especially as the data is self-reported and
diagnostic status could not be independently verified in the present
study.
Previous research has identified
higher levels of autistic-like traits being present within scientific
disciplines in which computer science students and employees are
included (Baron-Cohen et al. 2001; Billington et al. 2007; Ruzich et al. 2015b).
This study is the first to specify a direct relationship between higher
levels of autistic-like traits and advanced digital skills. In addition
to being a pre-requisite for committing cyber-dependent crimes, these
skills are essential for the cyber security industry which will have an
estimated 3.5 million unfulfilled jobs by 2021 (Morgan 2018).
This study suggests that targeting groups high in autistic-like traits
would be a beneficial strategy to meet this employment need. Given the
employment difficulties that can be faced by members of the autistic
community (Buescher et al. 2014; Knapp et al. 2009; see also Gotham et al. 2015; Hendricks 2010; Howlin 2013; Levy and Perry 2011; National Autistic Society 2016; Taylor et al. 2015; Shattuck et al. 2012)
and that around 46% of autistic adults who are employed are either
over-educated or exceed the skill level needed for the roles they are in
Baldwin et al. (2014), targeting the autistic community for cyber security employment may be particularly beneficial.
Notwithstanding
the limitations described above, this may be particularly pertinent as
this study found that a diagnosis of autism was associated with reduced
cyber-dependent criminality. This would be consistent with perceptions
of autistic strengths of honesty and loyalty (de Schipper et al. 2016)—ideal
attributes within employment settings. Importantly, this is not to
suggest that all autistic people are good with technology, or that all
autistic people should seek employment within cyber security industries
(see Milton 2018).
Rather, this study highlights that in a particularly challenging
employment context, some members of the autistic community may be
ideally suited to such employment opportunities and emphasises the need
for employers to ensure that their recruitment methods and working
environments are autism-friendly and inclusive (see Hedley et al. 2017 for review).
The
direct link between autistic-like traits and cyber-dependent crime is
also consistent with previous research (Seigfried-Spellar et al. 2015) and may extend to a relationship with cyber-enabled crime (such as online fraud). Seigfried-Spellar et al. (2015)
explored relationships between autistic-like traits and cyber-deviancy
more broadly defined than cyber-dependent crime. Future research could
explore whether the level of autistic-like traits, mediated by advanced
digital skills, also relates to cyber-enabled crime, and whether there
are any direct effects that are specific to cyber-dependent crime.
Seigfried-Spellar et al. (2015)
and the present study were both cross-sectional studies. The mediation
of advanced digital skills between autistic-like traits and
cyber-dependent crime has been assumed in the present study, but this
could be best established in longitudinal research. Exploring prison
populations to identify if ‘traditional’ crime was related to
autistic-like traits found no differences between prisoners and the
general population (Underwood et al. 2016),
which may suggest that autistic-like traits are associated with
cybercrime specifically (that is, cyber-dependent crime and potentially
cyber-enabled crime).
Sex, age, non-verbal
IQ, explicit social cognition and perceived interpersonal support did
not significantly relate to cyber-dependent criminal activity, which
serves to highlight the salience of autistic-like traits. A potential
limitation is that explicit social cognition was assessed, but not
implicit social cognition. Based on the autism literature (Callenmark et
al. 2014; Dewey 1991; Frith and Happé 1999),
we would not necessarily expect difficulties with explicit social
cognition in groups with high autistic-like traits. Implicit social
cognition was also assessed by Callenmark et al. using interviews after
the IToSK. Such interviews, however, do not readily extend to the online
context and future research could explore any role of implicit social
cognition in cyber-dependent crime. However, recent accounts of implicit
social cognition have questioned whether such a system exists and
findings from such measures can better be attributed to general
attentional processes (Conway et al. 2017; Heyes 2014; Santiesteban et al. 2014, 2015, 2017).
Future
research should also focus on autistic communities as well as those
convicted of cyber-dependent and cyber-enabled crimes to further develop
our understanding of this area, an important aspect of which is the
potential strengths some members of the autistic community can bring to
cyber security employment.