Cookies?
Library Header Image
LSE Research Online LSE Library Services

Artificial intelligence and decision-making: the question of accountability

Gualdi, Francesco and Cordella, Antonio (2021) Artificial intelligence and decision-making: the question of accountability. In: Bui, Tung X., (ed.) Proceedings of the 54th Annual Hawaii International Conference on System Sciences, HICSS 2021. Proceedings of the Annual Hawaii International Conference on System Sciences. IEEE Computer Society Press, Virtual, Online, 2297 - 2306. ISBN 9780998133140

[img] Text (Artificial intelligence and decision-making: the question of accountability) - Published Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (414kB)

Identification Number: 10.24251/HICSS.2021.281

Abstract

Public sector organizations literature has addressed the influence of AI on decision-making process, looking mainly at rationalization and efficiency. However, recent adoptions of AI have been challenged because of their discriminatory nature. As a result, questions emerged on the accountability of AI supported decision-making processes in the public sector. This research sheds light on how AI transforms decision-making processes in the public sector and hence on their accountability. The paper illustrates that AI adoptions lead to the emergency of techno-legal entanglements - assemblages - which might impact upon AI accountability. Building on the findings of some of the most controversial and discussed cases of AI adoption in the public sector - COMPAS in the US and UKVI in the UK - the paper makes the case for a new approach to AI supported public sector decision-making accountability.

Item Type: Book Section
Official URL: https://hicss.hawaii.edu/
Additional Information: © 2021 The Authors
Divisions: Management
Subjects: Q Science > QA Mathematics > QA76 Computer software
B Philosophy. Psychology. Religion > BJ Ethics
Date Deposited: 06 Jul 2021 12:51
Last Modified: 20 Sep 2021 02:29
URI: http://eprints.lse.ac.uk/id/eprint/110995

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

View more statistics