Title

How Much Automation Is Too Much? Keeping The Human Relevant In Knowledge Work

Keywords

Accounting profession; Artificial intelligence; Automation bias; Deskilling; Epistemology; Technology dominance

Abstract

With the rapid advances in data analytics, machine learning, and continuous monitoring along with other related advances in artificial intelligence-based technologies, our solution as researchers to many of today’s business problems increasingly becomes one of, ‘‘Can I fix the problem through automation?’’ However, as we find that artificial intelligence increasingly provides us with the power to replace knowledge workers with automated systems, rarely is the question asked, ‘‘Should we automate knowledge work?’’ There are a host of questions that should be addressed including (1) whether automation is the most effective solution, (2) if there are ethical dilemmas associated with replacing the human element, and (3) if there are societal implications of displacing large numbers of knowledge workers. The focus of this discussion is on understanding the impact of knowledge-based systems on human users’ knowledge acquisition and retention and outlining an alternative research strategy that centers more on transferring knowledge to the user during the work production process in order to maintain human expertise and relevance in professional decision making. Contemporary research still argues that human-computer collaboration may outperform either on their own; but, to limit the deskilling effect of knowledge-based systems and alternatively promote skill development, we call upon academic researchers to seek better ways to keep the human relevant in a broad range of knowledge work fields. Further, we suggest that expanding the philosophical discussions of the ethics of artificial intelligence-based technologies and the corollary impact on the rapid decline of the professions is necessary.

Publication Date

9-1-2018

Publication Title

Journal of Emerging Technologies in Accounting

Volume

15

Issue

2

Number of Pages

15-25

Document Type

Article

Personal Identifier

scopus

DOI Link

https://doi.org/10.2308/jeta-52311

Socpus ID

85070770316 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/85070770316

This document is currently not available here.

Share

COinS