Abstract
Going a step beyond decision automation discussed earlier, this chapter examines the growing role of algorithmic decision-making systems that are rapidly becoming the technological backbone of modern organizational functioning. Building on the idea of decision automation, this chapter takes a critical look at a diverse set of increasingly self-functioning applications designed to efficiently transform torrents of real-time and near-real-time time into decisions and actions that are not only executed more expediently but also, presumptively, more objectively. Given the seemingly unstoppable march of informational automation and the manner in which it is reshaping not only how work is done but also how lives are lived, the intent here is to offer a high-level overview of the Internet of Things–powered automation as seen from the perspective of some of the fundamental assumptions of algorithmic decision-making, especially those relating to bias, transparency and accountability considerations. And lastly, continuing with the theme of data-enabled creativity, a more substantive examination of the idea of ‘association’ is also undertaken, in the context of reimagined essence of data.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Interestingly, a variant of that idea actually existed in the United Kingdom until about 1950, where select universities, such as Oxford, had their own constituencies; those with a degree from one of those universities could vote both in their university constituency and their place of residence (in other words, elite had two votes). In the United States, numerous southern states as well as some northern states, such as New York and Connecticut, enacted various forms of voter literacy tests until all such tests were permanently outlawed by the US Congress in 1975.
- 2.
A typical non-machine learning-based system, such as the customer loyalty classification DSS algorithm discussed in Chap. 7, is compositionally fixed, which means that while input data may change, a scoring algorithm, expressed as a mathematical equation, remains unchanged (until it is redesigned or otherwise tweaked). In contrast to that, machine learning algorithms are compositionally dynamic, which means that the output producing mechanism (i.e. the composition of the hidden layers in Fig.8.1) continues to change with changing inputs.
- 3.
Though controversial at the time, primarily because of lack of clearly defined legal justification or precedent, the Nuremberg trials are now widely considered an important milestone towards establishment of a permanent international court for addressing crimes against humanity, such as genocide.
- 4.
Nuremberg Principle IV
- 5.
- 6.
The term ‘robber baron’ most likely originated in medieval Europe in reference to feudal lords who robbed merchant ships and travellers along the Rhine River; its modern use is attributed to a 1934 book by Matthew Josephson titled The Robber Barons.
- 7.
Personal and larger computers are considered ‘traditional’ computing devices, while the more recent mobile, wearable and other forms of data processing digital electronics are framed here as ‘non-traditional’.
- 8.
- 9.
This framing implicitly assumes a cross-sectional dataset in which an attribute, such as ‘age’, can take on different values across successive data records; in a longitudinal, or time-based dataset, however, a data record could represent time-based value, as illustrated by daily price of a particular stock. It is worth noting that a dataset can have a mixed, cross-sectional + longitudinal layout, as illustrated by a file containing daily stock prices for all New York Stock Exchange-listed companies for 2020.
- 10.
Broadly characterized, exploratory analyses seek to identify previously unknown patterns or associations, while confirmatory analyses are geared towards empirically testing of theoretically presumed associations (see the Conjectures and Refutations section in Chap. 3).
- 11.
The two most commonly used variants of correlation coefficients are Pearson’s product moment and Spearman’s rank. The former (which tends to be used more commonly, though sometimes incorrectly) is appropriate to use when both variables are continuous and normally distributed (which also means no outliers); the latter is appropriate when both variables are either skewed (i.e. not normally distributed) or ordinal (expressed as rank-ordered categories), and it is also robust with respect to the presence of outliers.
- 12.
Those are commonly referred to as ‘factors’ because factor analysis, a variable grouping statistical technique, is often used to construct statistically valid and reliable indicator composites.
- 13.
Formally defined, metavariable is an idea used in the study of semantics where it is framed as an element of a metalanguage (language used to describe expressions of another language) that can be used to refer to expressions in a logical language.
- 14.
An additional complication stems from advertising being a ‘slow moving’ variable, meaning that its level (as measure either in terms of spending or exposure) changes slowly, while sales levels fluctuate constantly.
- 15.
Formally, this phenomenon falls under a general umbrella of dual-processing theories of cognition, which recognize two types of information processing: conscious and subconscious (or non-conscious).
Reference
Zuboff, S. (2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power. New York: Public Affairs.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2021 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this chapter
Cite this chapter
Banasiewicz, A. (2021). In Machine We Trust. In: Organizational Learning in the Age of Data. EAI/Springer Innovations in Communication and Computing. Springer, Cham. https://doi.org/10.1007/978-3-030-74866-1_8
Download citation
DOI: https://doi.org/10.1007/978-3-030-74866-1_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-74865-4
Online ISBN: 978-3-030-74866-1
eBook Packages: EngineeringEngineering (R0)