Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance

Fabian Walke; Lars Bennek; Till J. Winkler

doi:10.1007/978-3-031-80122-8_8

Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance

Fabian Walke
, Lars Bennek
, Till J. Winkler^*

^*Corresponding author for this work

Department of Digitalization

FernUniversität in Hagen

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

Abstract

Explainability in artificial intelligence (AI) is crucial for ensuring transparency, accountability, and risk mitigation, thereby addressing digital responsibility, social, ethical and ecological aspects of information system usage. AI will be regulated in the European Union (EU) through the AI Act. This regulation introduces requirements for explainable AI (XAI). This paper examines which requirements for XAI are regulated and which metrics could be used for measuring compliance. For this purpose, legal texts from the European Parliament and Council were analyzed in order to ascertain XAI requirements. Additionally, XAI taxonomies and metrics were collected. The results reveal, that the AI Act provides abstract regulations for explainability, making it challenging to define specific metrics for achieving explainability. As a solution, we propose a socio-technical metric classification for measuring compliance. Further studies should analyze forthcoming explainability requirements to make AI verifiable and minimize risks arising from AI.

Original language	English
Title of host publication	Solutions and Technologies for Responsible Digitalization : Proceedings of the 18th International Conference on Wirtschaftsinformatik, Paderborn, Germany, 2023, Vol. 2
Editors	Daniel Beverungen, Christiane Lehrer, Matthias Trier
Number of pages	17
Place of Publication	Cham
Publisher	Springer
Publication date	2025
Pages	113-129
ISBN (Print)	9783031801211
ISBN (Electronic)	9783031801228
DOIs	https://doi.org/10.1007/978-3-031-80122-8_8
Publication status	Published - 2025
Event	18. Internationale Tagung Wirtschaftsinformatik. WI 2023 - Universität Paderborn, Paderborn, Germany Duration: 18 Sept 2023 → 21 Sept 2023 Conference number: 18 https://wi2023.de/

Conference

Conference	18. Internationale Tagung Wirtschaftsinformatik. WI 2023
Number	18
Location	Universität Paderborn
Country/Territory	Germany
City	Paderborn
Period	18/09/2023 → 21/09/2023
Internet address	https://wi2023.de/

Series	Lecture Notes in Information Systems and Organisation
Volume	75
ISSN	2195-4968

Keywords

XAI
Measurement
Assessment
Taxonomy
European union

Access to Document

10.1007/978-3-031-80122-8_8Licence: Unspecified

Handle.net

Persistent link

Cite this

Walke, F., Bennek, L., & Winkler, T. J. (2025). Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance. In D. Beverungen, C. Lehrer, & M. Trier (Eds.), Solutions and Technologies for Responsible Digitalization: Proceedings of the 18th International Conference on Wirtschaftsinformatik, Paderborn, Germany, 2023, Vol. 2 (pp. 113-129). Springer. https://doi.org/10.1007/978-3-031-80122-8_8

Walke, Fabian ; Bennek, Lars ; Winkler, Till J. / Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance. Solutions and Technologies for Responsible Digitalization: Proceedings of the 18th International Conference on Wirtschaftsinformatik, Paderborn, Germany, 2023, Vol. 2. editor / Daniel Beverungen ; Christiane Lehrer ; Matthias Trier. Cham : Springer, 2025. pp. 113-129 (Lecture Notes in Information Systems and Organisation, Vol. 75).

@inproceedings{e15c4a43edc64fcc83c1d506d95bf799,

title = "Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance",

abstract = "Explainability in artificial intelligence (AI) is crucial for ensuring transparency, accountability, and risk mitigation, thereby addressing digital responsibility, social, ethical and ecological aspects of information system usage. AI will be regulated in the European Union (EU) through the AI Act. This regulation introduces requirements for explainable AI (XAI). This paper examines which requirements for XAI are regulated and which metrics could be used for measuring compliance. For this purpose, legal texts from the European Parliament and Council were analyzed in order to ascertain XAI requirements. Additionally, XAI taxonomies and metrics were collected. The results reveal, that the AI Act provides abstract regulations for explainability, making it challenging to define specific metrics for achieving explainability. As a solution, we propose a socio-technical metric classification for measuring compliance. Further studies should analyze forthcoming explainability requirements to make AI verifiable and minimize risks arising from AI.",

keywords = "XAI, Measurement, Assessment, Taxonomy, European union, XAI, Measurement, Assessment, Taxonomy, European union",

author = "Fabian Walke and Lars Bennek and Winkler, \{Till J.\}",

year = "2025",

doi = "10.1007/978-3-031-80122-8\_8",

language = "English",

isbn = "9783031801211",

series = "Lecture Notes in Information Systems and Organisation",

publisher = "Springer",

pages = "113--129",

editor = "Daniel Beverungen and Christiane Lehrer and Matthias Trier",

booktitle = "Solutions and Technologies for Responsible Digitalization",

address = "Germany",

note = "18. Internationale Tagung Wirtschaftsinformatik. WI 2023, WI2023 ; Conference date: 18-09-2023 Through 21-09-2023",

url = "https://wi2023.de/",

}

Walke, F, Bennek, L & Winkler, TJ 2025, Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance. in D Beverungen, C Lehrer & M Trier (eds), Solutions and Technologies for Responsible Digitalization: Proceedings of the 18th International Conference on Wirtschaftsinformatik, Paderborn, Germany, 2023, Vol. 2. Springer, Cham, Lecture Notes in Information Systems and Organisation, vol. 75, pp. 113-129, 18. Internationale Tagung Wirtschaftsinformatik. WI 2023, Paderborn, North Rhine-Westphalia, Germany, 18/09/2023. https://doi.org/10.1007/978-3-031-80122-8_8

Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance. / Walke, Fabian; Bennek, Lars; Winkler, Till J.
Solutions and Technologies for Responsible Digitalization: Proceedings of the 18th International Conference on Wirtschaftsinformatik, Paderborn, Germany, 2023, Vol. 2. ed. / Daniel Beverungen; Christiane Lehrer; Matthias Trier. Cham: Springer, 2025. p. 113-129 (Lecture Notes in Information Systems and Organisation, Vol. 75).

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

TY - GEN

T1 - Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance

AU - Walke, Fabian

AU - Bennek, Lars

AU - Winkler, Till J.

N1 - Conference code: 18

PY - 2025

Y1 - 2025

N2 - Explainability in artificial intelligence (AI) is crucial for ensuring transparency, accountability, and risk mitigation, thereby addressing digital responsibility, social, ethical and ecological aspects of information system usage. AI will be regulated in the European Union (EU) through the AI Act. This regulation introduces requirements for explainable AI (XAI). This paper examines which requirements for XAI are regulated and which metrics could be used for measuring compliance. For this purpose, legal texts from the European Parliament and Council were analyzed in order to ascertain XAI requirements. Additionally, XAI taxonomies and metrics were collected. The results reveal, that the AI Act provides abstract regulations for explainability, making it challenging to define specific metrics for achieving explainability. As a solution, we propose a socio-technical metric classification for measuring compliance. Further studies should analyze forthcoming explainability requirements to make AI verifiable and minimize risks arising from AI.

AB - Explainability in artificial intelligence (AI) is crucial for ensuring transparency, accountability, and risk mitigation, thereby addressing digital responsibility, social, ethical and ecological aspects of information system usage. AI will be regulated in the European Union (EU) through the AI Act. This regulation introduces requirements for explainable AI (XAI). This paper examines which requirements for XAI are regulated and which metrics could be used for measuring compliance. For this purpose, legal texts from the European Parliament and Council were analyzed in order to ascertain XAI requirements. Additionally, XAI taxonomies and metrics were collected. The results reveal, that the AI Act provides abstract regulations for explainability, making it challenging to define specific metrics for achieving explainability. As a solution, we propose a socio-technical metric classification for measuring compliance. Further studies should analyze forthcoming explainability requirements to make AI verifiable and minimize risks arising from AI.

KW - XAI

KW - Measurement

KW - Assessment

KW - Taxonomy

KW - European union

KW - XAI

KW - Measurement

KW - Assessment

KW - Taxonomy

KW - European union

U2 - 10.1007/978-3-031-80122-8_8

DO - 10.1007/978-3-031-80122-8_8

M3 - Article in proceedings

SN - 9783031801211

T3 - Lecture Notes in Information Systems and Organisation

SP - 113

EP - 129

BT - Solutions and Technologies for Responsible Digitalization

A2 - Beverungen, Daniel

A2 - Lehrer, Christiane

A2 - Trier, Matthias

PB - Springer

CY - Cham

T2 - 18. Internationale Tagung Wirtschaftsinformatik. WI 2023

Y2 - 18 September 2023 through 21 September 2023

ER -

Walke F, Bennek L, Winkler TJ. Artificial Intelligence Explainability Requirements of the AI Act and Metrics for Measuring Compliance. In Beverungen D, Lehrer C, Trier M, editors, Solutions and Technologies for Responsible Digitalization: Proceedings of the 18th International Conference on Wirtschaftsinformatik, Paderborn, Germany, 2023, Vol. 2. Cham: Springer. 2025. p. 113-129. (Lecture Notes in Information Systems and Organisation, Vol. 75). doi: 10.1007/978-3-031-80122-8_8