Towards a Taxonomy of AI Risks in the Health Domain

Delaram Golpayegani; Joshua Hovsha; Leon W.S. Rossmaier; Rana Saniei; Jana Mišić

doi:10.1109/TransAI54797.2022.00007

Towards a Taxonomy of AI Risks in the Health Domain

Delaram Golpayegani, Joshua Hovsha, Leon W.S. Rossmaier, Rana Saniei, Jana Mišić

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

2 Citations (Scopus)

263 Downloads (Pure)

Abstract

The adoption of AI in the health sector has its share of benefits and harms to various stakeholder groups and entities. There are critical risks involved in using AI systems in the health domain; risks that can have severe, irreversible, and life-changing impacts on people's lives. With the development of innovative AI-based applications in the medical and healthcare sectors, new types of risks emerge. To benefit from novel AI applications in this domain, the risks need to be managed in order to protect the fundamental interests and rights of those affected. This will increase the level to which these systems become ethically acceptable, legally permissible, and socially sustainable. In this paper, we first discuss the necessity of AI risk management in the health domain from the ethical, legal, and societal perspectives. We then present a taxonomy of risks associated with the use of AI systems in the health domain called HART, accessible online at https://w3id.org/hart. HART mirrors the risks of a variety of different real-world incidents caused by use of AI in the health sector. Lastly, we discuss the implications of the taxonomy for different stakeholder groups and further research.

Original language	English
Title of host publication	Proceedings - 2022 4th International Conference on Transdisciplinary AI, TransAI 2022
Place of Publication	Piscataway, NJ
Publisher	IEEE
Pages	1-8
Number of pages	8
ISBN (Electronic)	978-1-6654-7184-8
ISBN (Print)	978-1-6654-7185-5
DOIs	https://doi.org/10.1109/TransAI54797.2022.00007
Publication status	Published - 2022
Event	4th International Conference on Transdisciplinary AI, TransAI 2022 - Laguna Hills, United States Duration: 19 Sept 2022 → 21 Sept 2022 Conference number: 4

Conference

Conference	4th International Conference on Transdisciplinary AI, TransAI 2022
Abbreviated title	TransAI 2022
Country/Territory	United States
City	Laguna Hills
Period	19/09/22 → 21/09/22

Keywords

AI public policy
AI regulation
AI systems
Ethics of AI
Health
Risk
Taxonomy
2023 OA procedure

Access to Document

10.1109/TransAI54797.2022.00007

ArticleFinal published version, 769 KBLicence: Taverne

Cite this

@inproceedings{5e2373136b1f438ebeea851b979a5668,

title = "Towards a Taxonomy of AI Risks in the Health Domain",

abstract = "The adoption of AI in the health sector has its share of benefits and harms to various stakeholder groups and entities. There are critical risks involved in using AI systems in the health domain; risks that can have severe, irreversible, and life-changing impacts on people's lives. With the development of innovative AI-based applications in the medical and healthcare sectors, new types of risks emerge. To benefit from novel AI applications in this domain, the risks need to be managed in order to protect the fundamental interests and rights of those affected. This will increase the level to which these systems become ethically acceptable, legally permissible, and socially sustainable. In this paper, we first discuss the necessity of AI risk management in the health domain from the ethical, legal, and societal perspectives. We then present a taxonomy of risks associated with the use of AI systems in the health domain called HART, accessible online at https://w3id.org/hart. HART mirrors the risks of a variety of different real-world incidents caused by use of AI in the health sector. Lastly, we discuss the implications of the taxonomy for different stakeholder groups and further research.",

keywords = "AI public policy, AI regulation, AI systems, Ethics of AI, Health, Risk, Taxonomy, 2023 OA procedure",

author = "Delaram Golpayegani and Joshua Hovsha and Rossmaier, {Leon W.S.} and Rana Saniei and Jana Mi{\v s}i{\'c}",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 4th International Conference on Transdisciplinary AI, TransAI 2022, TransAI 2022 ; Conference date: 19-09-2022 Through 21-09-2022",

year = "2022",

doi = "10.1109/TransAI54797.2022.00007",

language = "English",

isbn = "978-1-6654-7185-5",

pages = "1--8",

booktitle = "Proceedings - 2022 4th International Conference on Transdisciplinary AI, TransAI 2022",

publisher = "IEEE",

address = "United States",

}

Golpayegani, D, Hovsha, J, Rossmaier, LWS, Saniei, R & Mišić, J 2022, Towards a Taxonomy of AI Risks in the Health Domain. in Proceedings - 2022 4th International Conference on Transdisciplinary AI, TransAI 2022. IEEE, Piscataway, NJ, pp. 1-8, 4th International Conference on Transdisciplinary AI, TransAI 2022, Laguna Hills, California, United States, 19/09/22. https://doi.org/10.1109/TransAI54797.2022.00007

TY - GEN

T1 - Towards a Taxonomy of AI Risks in the Health Domain

AU - Golpayegani, Delaram

AU - Hovsha, Joshua

AU - Rossmaier, Leon W.S.

AU - Saniei, Rana

AU - Mišić, Jana

N1 - Conference code: 4

PY - 2022

Y1 - 2022

N2 - The adoption of AI in the health sector has its share of benefits and harms to various stakeholder groups and entities. There are critical risks involved in using AI systems in the health domain; risks that can have severe, irreversible, and life-changing impacts on people's lives. With the development of innovative AI-based applications in the medical and healthcare sectors, new types of risks emerge. To benefit from novel AI applications in this domain, the risks need to be managed in order to protect the fundamental interests and rights of those affected. This will increase the level to which these systems become ethically acceptable, legally permissible, and socially sustainable. In this paper, we first discuss the necessity of AI risk management in the health domain from the ethical, legal, and societal perspectives. We then present a taxonomy of risks associated with the use of AI systems in the health domain called HART, accessible online at https://w3id.org/hart. HART mirrors the risks of a variety of different real-world incidents caused by use of AI in the health sector. Lastly, we discuss the implications of the taxonomy for different stakeholder groups and further research.

AB - The adoption of AI in the health sector has its share of benefits and harms to various stakeholder groups and entities. There are critical risks involved in using AI systems in the health domain; risks that can have severe, irreversible, and life-changing impacts on people's lives. With the development of innovative AI-based applications in the medical and healthcare sectors, new types of risks emerge. To benefit from novel AI applications in this domain, the risks need to be managed in order to protect the fundamental interests and rights of those affected. This will increase the level to which these systems become ethically acceptable, legally permissible, and socially sustainable. In this paper, we first discuss the necessity of AI risk management in the health domain from the ethical, legal, and societal perspectives. We then present a taxonomy of risks associated with the use of AI systems in the health domain called HART, accessible online at https://w3id.org/hart. HART mirrors the risks of a variety of different real-world incidents caused by use of AI in the health sector. Lastly, we discuss the implications of the taxonomy for different stakeholder groups and further research.

KW - AI public policy

KW - AI regulation

KW - AI systems

KW - Ethics of AI

KW - Health

KW - Risk

KW - Taxonomy

KW - 2023 OA procedure

UR - http://www.scopus.com/inward/record.url?scp=85143407649&partnerID=8YFLogxK

U2 - 10.1109/TransAI54797.2022.00007

DO - 10.1109/TransAI54797.2022.00007

M3 - Conference contribution

AN - SCOPUS:85143407649

SN - 978-1-6654-7185-5

SP - 1

EP - 8

BT - Proceedings - 2022 4th International Conference on Transdisciplinary AI, TransAI 2022

PB - IEEE

CY - Piscataway, NJ

T2 - 4th International Conference on Transdisciplinary AI, TransAI 2022

Y2 - 19 September 2022 through 21 September 2022

ER -

Towards a Taxonomy of AI Risks in the Health Domain

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this