Effects of Fairness and Explanation on Trust in Ethical AI

Alessa Angerschmid; Kevin Theuermann; Andreas Holzinger; Fang Chen; Jianlong Zhou

doi:10.1007/978-3-031-14463-9_4

Effects of Fairness and Explanation on Trust in Ethical AI

Alessa Angerschmid, Kevin Theuermann, Andreas Holzinger, Fang Chen, Jianlong Zhou^*

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Abstract

AI ethics has been a much discussed topic in recent years. Fairness and explainability are two important ethical principles for trustworthy AI. In this paper, the impact of AI explainability and fairness on user trust in AI-assisted decisions is investigated. For this purpose, a user study was conducted simulating AI-assisted decision making in a health insurance scenario. The study results demonstrated that fairness only affects user trust when the fairness level is low, with a low fairness level reducing user trust. However, adding explanations helped users increase their trust in AI-assisted decision making. The results show that the use of AI explanations and fairness statements in AI applications is complex: we need to consider not only the type of explanations, but also the level of fairness introduced. This is a strong motivation for further work.

Original language	English
Title of host publication	Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings
Editors	Andreas Holzinger, Andreas Holzinger, Andreas Holzinger, Peter Kieseberg, A Min Tjoa, Edgar Weippl, Edgar Weippl
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	51-67
Number of pages	17
ISBN (Print)	9783031144622
DOIs	https://doi.org/10.1007/978-3-031-14463-9_4
Publication status	Published - 2022
Event	6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference for Machine Learning and Knowledge Extraction, held in conjunction with the 17th International Conference on Availability, Reliability and Security: ARES 2022 - Vienna, Austria Duration: 23 Aug 2022 → 26 Aug 2022

Publication series

Name	Lecture Notes in Computer Science
Volume	13480 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference for Machine Learning and Knowledge Extraction, held in conjunction with the 17th International Conference on Availability, Reliability and Security
Country/Territory	Austria
City	Vienna
Period	23/08/22 → 26/08/22

Keywords

AI ethics
AI explanation
AI fairness
Trust

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-14463-9_4

Cite this

Angerschmid, A., Theuermann, K., Holzinger, A., Chen, F., & Zhou, J. (2022). Effects of Fairness and Explanation on Trust in Ethical AI. In A. Holzinger, A. Holzinger, A. Holzinger, P. Kieseberg, A. M. Tjoa, E. Weippl, & E. Weippl (Eds.), Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings (pp. 51-67). (Lecture Notes in Computer Science ; Vol. 13480 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-14463-9_4

Effects of Fairness and Explanation on Trust in Ethical AI. / Angerschmid, Alessa; Theuermann, Kevin; Holzinger, Andreas et al.
Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings. ed. / Andreas Holzinger; Andreas Holzinger; Andreas Holzinger; Peter Kieseberg; A Min Tjoa; Edgar Weippl; Edgar Weippl. Springer Science and Business Media Deutschland GmbH, 2022. p. 51-67 (Lecture Notes in Computer Science ; Vol. 13480 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Angerschmid, A, Theuermann, K, Holzinger, A, Chen, F & Zhou, J 2022, Effects of Fairness and Explanation on Trust in Ethical AI. in A Holzinger, A Holzinger, A Holzinger, P Kieseberg, AM Tjoa, E Weippl & E Weippl (eds), Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings. Lecture Notes in Computer Science , vol. 13480 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 51-67, 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference for Machine Learning and Knowledge Extraction, held in conjunction with the 17th International Conference on Availability, Reliability and Security, Vienna, Austria, 23/08/22. https://doi.org/10.1007/978-3-031-14463-9_4

Angerschmid A, Theuermann K, Holzinger A, Chen F, Zhou J. Effects of Fairness and Explanation on Trust in Ethical AI. In Holzinger A, Holzinger A, Holzinger A, Kieseberg P, Tjoa AM, Weippl E, Weippl E, editors, Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 51-67. (Lecture Notes in Computer Science ). doi: 10.1007/978-3-031-14463-9_4

Angerschmid, Alessa ; Theuermann, Kevin ; Holzinger, Andreas et al. / Effects of Fairness and Explanation on Trust in Ethical AI. Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings. editor / Andreas Holzinger ; Andreas Holzinger ; Andreas Holzinger ; Peter Kieseberg ; A Min Tjoa ; Edgar Weippl ; Edgar Weippl. Springer Science and Business Media Deutschland GmbH, 2022. pp. 51-67 (Lecture Notes in Computer Science ).

@inproceedings{1303378326b04748b4388664e86dc3e4,

title = "Effects of Fairness and Explanation on Trust in Ethical AI",

abstract = "AI ethics has been a much discussed topic in recent years. Fairness and explainability are two important ethical principles for trustworthy AI. In this paper, the impact of AI explainability and fairness on user trust in AI-assisted decisions is investigated. For this purpose, a user study was conducted simulating AI-assisted decision making in a health insurance scenario. The study results demonstrated that fairness only affects user trust when the fairness level is low, with a low fairness level reducing user trust. However, adding explanations helped users increase their trust in AI-assisted decision making. The results show that the use of AI explanations and fairness statements in AI applications is complex: we need to consider not only the type of explanations, but also the level of fairness introduced. This is a strong motivation for further work.",

keywords = "AI ethics, AI explanation, AI fairness, Trust",

author = "Alessa Angerschmid and Kevin Theuermann and Andreas Holzinger and Fang Chen and Jianlong Zhou",

note = "Funding Information: Acknowledgements. This work does not raise any ethical issues. Parts of this work have been funded by the Austrian Science Fund (FWF), Project: P-32554 explainable Artificial Intelligence; and by the Australian UTS STEM-HASS Strategic Research Fund 2021. Publisher Copyright: {\textcopyright} 2022, IFIP International Federation for Information Processing.; 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference for Machine Learning and Knowledge Extraction, held in conjunction with the 17th International Conference on Availability, Reliability and Security : ARES 2022 ; Conference date: 23-08-2022 Through 26-08-2022",

year = "2022",

doi = "10.1007/978-3-031-14463-9_4",

language = "English",

isbn = "9783031144622",

series = "Lecture Notes in Computer Science ",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "51--67",

editor = "Andreas Holzinger and Andreas Holzinger and Andreas Holzinger and Peter Kieseberg and Tjoa, {A Min} and Edgar Weippl and Edgar Weippl",

booktitle = "Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Effects of Fairness and Explanation on Trust in Ethical AI

AU - Angerschmid, Alessa

AU - Theuermann, Kevin

AU - Holzinger, Andreas

AU - Chen, Fang

AU - Zhou, Jianlong

N1 - Funding Information: Acknowledgements. This work does not raise any ethical issues. Parts of this work have been funded by the Austrian Science Fund (FWF), Project: P-32554 explainable Artificial Intelligence; and by the Australian UTS STEM-HASS Strategic Research Fund 2021. Publisher Copyright: © 2022, IFIP International Federation for Information Processing.

PY - 2022

Y1 - 2022

N2 - AI ethics has been a much discussed topic in recent years. Fairness and explainability are two important ethical principles for trustworthy AI. In this paper, the impact of AI explainability and fairness on user trust in AI-assisted decisions is investigated. For this purpose, a user study was conducted simulating AI-assisted decision making in a health insurance scenario. The study results demonstrated that fairness only affects user trust when the fairness level is low, with a low fairness level reducing user trust. However, adding explanations helped users increase their trust in AI-assisted decision making. The results show that the use of AI explanations and fairness statements in AI applications is complex: we need to consider not only the type of explanations, but also the level of fairness introduced. This is a strong motivation for further work.

AB - AI ethics has been a much discussed topic in recent years. Fairness and explainability are two important ethical principles for trustworthy AI. In this paper, the impact of AI explainability and fairness on user trust in AI-assisted decisions is investigated. For this purpose, a user study was conducted simulating AI-assisted decision making in a health insurance scenario. The study results demonstrated that fairness only affects user trust when the fairness level is low, with a low fairness level reducing user trust. However, adding explanations helped users increase their trust in AI-assisted decision making. The results show that the use of AI explanations and fairness statements in AI applications is complex: we need to consider not only the type of explanations, but also the level of fairness introduced. This is a strong motivation for further work.

KW - AI ethics

KW - AI explanation

KW - AI fairness

KW - Trust

UR - http://www.scopus.com/inward/record.url?scp=85136995893&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-14463-9_4

DO - 10.1007/978-3-031-14463-9_4

M3 - Conference paper

AN - SCOPUS:85136995893

SN - 9783031144622

T3 - Lecture Notes in Computer Science

SP - 51

EP - 67

BT - Machine Learning and Knowledge Extraction - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference, CD-MAKE 2022, Proceedings

A2 - Holzinger, Andreas

A2 - Kieseberg, Peter

A2 - Tjoa, A Min

A2 - Weippl, Edgar

PB - Springer Science and Business Media Deutschland GmbH

T2 - 6th IFIP TC 5, TC 12, WG 8.4, WG 8.9, WG 12.9 International Cross-Domain Conference for Machine Learning and Knowledge Extraction, held in conjunction with the 17th International Conference on Availability, Reliability and Security

Y2 - 23 August 2022 through 26 August 2022

ER -