The ghost in the legal machine: algorithmic governmentality, economy, and the practice of law
| Pages | 16-31 |
| DOI | https://doi.org/10.1108/JICES-09-2016-0038 |
| Date | 12 March 2018 |
| Published date | 12 March 2018 |
| Author | Adam Harkens |
| Subject Matter | Information & knowledge management,Information management & governance,Information & communications technology |
The ghost in the legal machine:
algorithmic governmentality,
economy, and the practice of law
Adam Harkens
School of Law, Queen’s University Belfast, Belfast, UK
Abstract
Purpose –This paper aims to investigate algorithmic governmentality –as proposed by Antoinette
Rouvroy –specifically in relation to law. It seeks to show how algorithmic profiling can be particularly
attractive for thosein legal practice, given restraints on time and resources.It deviates from Rouvroy in two
ways. First, it argues that algorithmic governmentality does not contrast with neoliberal modes of
government in that it allows indirect rule through economic calculations. Second, it argues that critique of
such systemsis possible,especially if the creative nature of law can be harnessed effectively.
Design/methodology/approach –This is a conceptual paper, with a theory-based approach, that is
intended to explore relevant issues related to algorithmic governmentality as a basis for future empirical
research.It builds on governmentality and socio-legal studies, as well as researchon algorithmic practices and
some documentaryanalysis of reports and public-facingmarketing of relevant technologies.
Findings –This paper provides insights on how algorithmic knowledge is collected, constructed and
applied in differentsituations. It provides examples of how algorithms are currentlyused and how trends are
developing.It demonstrates how such uses can be informed by socio-politicaland economic rationalities.
Research limitations/implications –Further empirical research is required to test the theoretical
findings.
Originality/value –This paper takesup Rouvroy’s question of whether we are at the end(s) of critique and
seeks to identify where such critique can be made possible. It also highlights the importance of
acknowledgingthe role of political rationalities in informing the activityof algorithmic assemblages.
Keywords Law, Economy, Socio-technical systems, Information and power,
Algorithmic governmentality, Legal practice
Paper type Conceptual paper
Introduction
Algorithmic governmentalityis a term coined by Antoinette Rouvroy (2012) to describe new
regimes of power brought forward by the “computational turn”and the ever-increasing
prevalence of algorithms in daily life. Popular applications includethe structuring of social
media timelines, provision of tailored music and film recommendations and filtering and
dissemination of news articles. Each of these applications shares a common theme in that
This research has been carried out as part of the Leverhulme Interdisciplinary Network on
Cybersecurity and Society (LINCS) at Queen’s University Belfast. Funding for this project has been
jointly provided by the Leverhulme Trust and Queen’s University Belfast, where the original
programme design and bid originated. The author is very grateful for the advice given by Professors
John Morison and Daniel Crookes, and Dr Muiris McMacCarthaigh on earlier versions of this paper;
Simon Rogerson and the reviewers for their extremely helpful comments; and Anthony Behan and
Georgios Glouftsios for their valuable input.
JICES
16,1
16
Received25 September 2016
Revised18 April 2017
Accepted18 April 2017
Journalof Information,
Communicationand Ethics in
Society
Vol.16 No. 1, 2018
pp. 16-31
© Emerald Publishing Limited
1477-996X
DOI 10.1108/JICES-09-2016-0038
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1477-996X.htm
they bring a structured order to the chaos of the online world –so that informationbecomes
more easily digestibleand manageable for the everyday user.
However, algorithms also enable the profiling of these same users through data-mining
practices, thus allowing a form of financial risk management aiming at “minimizing the
uncertainty associated to human agency”(p.10). In this manner then, the chaos of worldly
potentialities also becomes ordered and more easily digestible for those charged with the
tasks of governing and decision-making,regardless of whether they are accurate (and if not,
this threatens potentially direconsequences for the profiled individual). Algorithms are not
fussy about whom theywork for.
This paper focuses on a specificdefinition of “governmentality”given by Foucault and
subsequent scholars. I will also attempt to demarcate points of differentiation between
Rouvroy’s thought,and my own, so as to prevent any confusion.
Foucault (2008, p. 186) describes the “art of government”as the way in which one
“conducts the conduct of men”or endeavours to “shape, guide or direct”the conduct of
others (Rose, 1999,p.15).Miller and Rose (2008, pp. 15-16) split this into two aspects: the
first being rationalities, mentalities or “styles of thinking, ways of rendering reality
thinkable in such a way that it [becomes] amenable to calculationand programming”, which
are rational in the sense that they privilege “systematic ways of thinking”, such as the
languages of economics and management,“over symbolic, mythical or poetic modes”(Dean,
2010, p. 25); and the second, of human technologies or “assemblages of persons, techniques,
institutions, instruments for the conducting of conduct”. This article places algorithms
firmly within the second aspect, as instrumentsof government, which hold the potential for
distinctive effect based upon governmental rationality with which they are used, and the
assemblages they are situatedwithin.
It does so by looking at one particularconcept that has become more prominent in recent
times, and relies on algorithms to work effectively: big data. Big data is at the forefront of
aforementioned algorithmic risk management techniques, which enact what Antoinette
Rouvroy (2012, p. 143) refersto as data-behaviourism, or new ways of “producing knowledge
[about future events] withoutconsidering the subject’s psychological motivations, speeches
or narratives, but rather relyingon data”.
Once this knowledge is collatedand organised, it can be acted upon due to the belief in its
“objective”value. Certain areas have become particularly amenable to such practices,
including online advertising, personal insurance, health [as evidenced by both the concepts
of quantified self and genomics, whereby datascientists such as Jeff Hammerbacher seek to
make medicine an “informationgame”(Lohr, 2015, pp. 163-164)] and crucially to this specific
article, law and legal knowledge.
Nevertheless, this paper takes a step away from Rouvroy’s contention that algorithmic
governmentality “contrasts with what we know about a neoliberal mode of government
which produces the subjects it needs”(p. 8). On the contrary,algorithms represent the ideal
tools for a neoliberal form of economic management, entailing what Miller and Rose (2008,
p. 16) refer to as government at a distance.
In addition, while Rouvroy is not interested in “whether algorithms produce ‘valid’
predictions or not”(p. 143), this article argues that it is crucial to make any inaccuracies
within algorithmic systems clear to form the basis of critique. As algorithms become more
prominent in legal systems,the creative nature of law is crucial in forming these critiques, so
that the ghost in the legal machine (meaning here the discrete artificial “intelligence”
attributed to algorithms),is taken with a large pinch of salt –and the unintentional “lock-in”
of socio-legal normscan be prevented.
Ghost in the
legal machine
17
To continue reading
Request your trial