IEEE Technology and Society Magazine - June 2019 - 42

software to address a complex issue or solve a specific
problem by automating what would otherwise require
human mental effort.2
Specialized AI is best seen as an extension of more
traditional practices such as software engineering, IT
systems design, database management and data science which deploys a range of AI techniques to automate the search for solutions to problems that currently
require substantial human mental labor and skill. Much
of the current excitement around AI is focused on "deep
learning" machine learning techniques that use manylayered "deep" neural networks that can find complex
patterns in large datasets ("big data"). Far from artificial
sentience, consciousness or general intelligence, we
could consider this as enthusiasm for "statistics on steroids." Commercial and governmental institutions have
long used statistics to develop representations of the
world that can inform future actions and policies. In this
sense, the AI revolution is really a continuation, and
massive acceleration, of much longer and older trends
of datafication and computerization. What is new and
unprecedented is the sheer volume of data, the speed
at which it can now be effectively processed, the sophistication of the analysis of that data, the degree of automation and the consequent lack of direct human
oversight that is possible.
As data-driven organizational management - led
by big data, machine learning and AI techniques -
continues to accelerate, and more processes are automated, there are growing concerns over the social and
ethical implications of this transformation. Machine
ethics is concerned with how autonomous systems
can be imbued with ethical values. "AI ethics" considers both designing AI to explicitly recognize and solve
ethical problems, and the implicit values and ethics of
implementing various AI applications and making
automated decisions with ethical consequences. This
paper will consider the latter, implicit view that corresponds to what is sometimes called "robot ethics," to
distinguish it from explicit "machine ethics" [4]. Ideally, the explicit ethics, implicit ethics, and the embedding and regulation of the system in society should all
align [5].
The outputs of predictive policing algorithms clearly
have ethical consequences, even if the systems under
consideration do not try to design systems for explicit
ethical reasoning. In the predictive policing systems
under consideration, there is little or no effort to design
the systems to frame their analysis or results as ethical
2
Some theorists have speculated about the possibility or consequences
of an artificial general intelligence (AGI) which might be able to learn with
little or no direct instruction from humans, and in some sense recognize
problems on its own that are in need of solution, and then adapt itself to
solve them. AGI is not technologically feasible for the foreseeable future,
and as such it will not be given much consideration here.

42

decisions or perform ethical analyses. What is of concern to the public, and in this paper, is how well the systems are designed, and the ethical implications of
introducing them into police practices.
There is a growing body of research examining the
ways in which data-driven algorithms are being used in
an increasing number of critical decision processes,
often with little or no accountability [6]-[9], and sometimes with little or no real understanding of how they
function in the real world or why they reach the results
they do in particular cases [10]-[12]. Consequently,
there are many ways for such systems to "go wrong."
Sometimes this is due to a well-intentioned but mathematically naive understanding of how such systems
work. This includes the failure to understand how
statistical outliers may be mishandled or misrepresented, or how historical data patterns can be selfreinforcing - such as denying credit and charging
higher interest rates to poorer individuals and communities, thus systematically denying them opportunities to escape poverty. Sometimes this is due to the
intended desire to transfer responsibility and blame
to an automated process, and relieve human agents
of their responsibility. And sometimes there may be
malevolent motives behind using data in obviously
discriminatory ways - such as purging voter rolls to
deny eligible voters to an opposing political party. But
these are ultimately "narrow" views of AI ethics, which
look to improving accuracy and performance of the
technology, while largely ignoring the context of use. It
has also been argued that the focus of AI ethics on
"solving" the bias problem is a distraction from other
and more important ethical and social issues [13].
Without discounting the value of such narrow approaches, this paper will examine the importance of taking a
broader ethical perspective on AI, and the problems
that will not be fixed through fairness, accountability
and transparency alone.

Two Approaches to AI Ethics
This paper aims to go beyond the ways in which data
and AI algorithms might be biased or unaccountable,
and consider the ethics of how AI systems are em bedded in social practices. Because AI ostensibly
automates various forms of human reasoning, consideration, and judgement, the accuracy or fairness of such
processes alone do not guarantee that their use will provide just, ethical, and socially desirable results. Rather,
careful attention must be paid to the ways in which the
implementation of such systems changes the practices
of those who use them. In order to redirect attention to
the bigger picture of the socio-technical embeddedness
of AI when considering ethics, the paper will formulate
two broad concepts of AI ethics, which will be named

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

JUNE 2019
IEEE Technology and Society Magazine - June 2019

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