IEEE Technology and Society Magazine - December 2019 - 60

evidence, and to engage in a civilized exchange of views
(rather than a playground slanging match) is, in these
days of polarized politics, a prosocial benefit.
On the other hand, automating the offside decision
system could bring a positive shift in postmatch debates
from discussions on contentious calls made by the referee to other important aspects of the game such as
team tactics and player performance analysis. This
would enhance the quality of debate, since it focuses
more on the game itself rather than fruitless arguments
over whether a match official's real time interpretation
of events coincided with an interpretation based on
endless slow-motion repeats, informed by experience
(of actually playing the game) or distorted, perhaps, by a
bias (in favor of a particular team).

Juridical Implications
To address the issue of acceptability, we note that the socalled platform revolution [23] has resulted in more and
more human-human interaction being conducted through
platforms, and these interactions are largely governed by
and regulated by rules embedded in the code that constitutes these platforms. Consequently, the increased digitization and reliance on algorithms for regulating
interactions brings a new dimension to rule enforcement
in different spheres. As formulated by Lessig [16], "code is
Law": code effectively determines the governance of the
Internet, and - as such - is capable of constraining an
individual's actions via technological means.
Regulation by algorithm is a promising mechanism
for rule or law enforcement that is progressively establishing itself not only in the private sector but also in the
public sector, e.g., for governance, policing, and statutory organizations (e.g., civil aviation authorities). Governments and public administrations increasingly rely on
software algorithms and technological tools in order to
define code-base rules, which are automatically executed (or enforced) by the underlying technology. As examples, the "No Fly List" in the U.S., relies on data mining
to assess potential threats to national security [24];
while algorithms have been used for judicial decision
making and to determine jail sentences or paroles [25].
However, the overall, and indeed eventual, impact of
algorithm-based regulation on different areas of human
enterprise, especially with regard to judicial decision
making (i.e., the administration of justice, law, or rules),
is difficult to evaluate because "we" are embedded
observers of a trajectory rather than external observers
of a state transition. Nevertheless, to legitimize a statutory authority to take critical decisions that could have
major implications on human life, in which those decisions are made primarily through numbers and statistics
and a literal, reductionist interpretation of the rules,
casts doubt on whether the system would be "humane"

60

in taking such critical decisions. For example, a system
that governs traffic regulations, indicting a person for a
minor traffic violation that could have occurred due to
the application of common sense, or simply lack of
knowledge, is counterintuitive to the human experience.
It also denies the "fudge factor," whereby people cheat
just a bit, but still see themselves as honest and honorable [29]; moreover the ability to break conventional
rules seems to be a necessary human capacity.10
Ultimately, the concern for the impact on judicial
administration is twofold. The first concern is that, similar to the "technology creep" observed in surveillance
capitalism [30], implants [31], and wearable computing
[32], the acceptable use of algorithmic definition in sport,
as a subdomain of entertainment, would predispose
acceptability in other walks of life where the outcome
may not actually be so clear cut. This could be, for example in medical triage, autonomous weapons systems (a
valid target is whatever the missile hits, by definition), or
jurisprudence (a person is - by algorithmic definition -
guilty if the algorithm outputs "s/he is guilty").
The second concern is that the rules clearly specify
aspects that are open to interpretation: "if, in the opinion
of the referee, the player is gaining an advantage." The
question here is actually about judgement of intention:
and algorithmic definition ignores intent. SmartFlag is
designed to deal with uncertainty, but uncertainty has
(at least) two dimensions: doubt and ambiguity [33].
Training data and analytics can help reduce doubt, but it
cannot help so much dealing with ambiguity. And that
ambiguity also exists in two dimensions: the intention of
the law-maker (e.g., to eliminate an "unfair" advantage)
and the intention of the (potential) lawbreaker (e.g., to
gain such an "unfair" advantage). Some offside decisions
are so fractional that the distances involved can hardly
have conferred an "unfair" advantage or made any difference to the outcome, except possibly to negate a
moment of exceptional skill.

Marginalization: Computer-Based Prediction
versus Human Prediction
In considering marginalization, we take into account
both marginalization of the data, as well as marginalization of the human.
10
For example, in U.K. traffic law, it is a matter of convention (as described
by the Highway Code) to give priority to an emergency vehicle. If in doing
so a driver commits a legal offense - for example entering a bus lane - s/
he is still liable for the fine in doing so. A human observer might apply common sense and forgiveness [26], but an automated system, following the
letter of the law (and easier to code) [27] rather than the "spirit" of the law
(what the regulators intended to achieve with their legislation, and much
harder to code), might not be so accommodating. Common sense, flexibility, and even disobedience in the following and enforcement of conventional rules are key features in the principled violation of policy [28]. Moreover
choosing to not apply a sanction for an observed indiscretion is also a way
of generating incentives and social capital. Note also that civil disobedience is a right, which in some countries is established in constitutional law.

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

DECEMBER 2019
IEEE Technology and Society Magazine - December 2019

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2019
