Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AuditMAI: Towards An Infrastructure for Continuous AI Auditing

Laura Waltersdorfer, Fajar J. Ekaputra, Tomasz Miksa, Marta Sabou

TL;DR

The paper addresses the lack of continuous and integrated AI auditing by proposing AuditMAI as a blueprint for an infrastructure enabling ongoing auditability. It grounds the framework in an explicit AI auditability definition with four key elements and derives four practical requirements from two industrial use cases. The authors describe a three-view architecture—Knowledge, Process, and Architecture—and outline a prototype AuditBox employing semantic-web techniques to integrate auditable artefacts. By combining knowledge management, systematic auditing processes, and architectural components, AuditMAI aims to reduce manual overhead and support preventive, regular AI risk assessments in real-world settings.

Abstract

Artificial Intelligence (AI) Auditability is a core requirement for achieving responsible AI system design. However, it is not yet a prominent design feature in current applications. Existing AI auditing tools typically lack integration features and remain as isolated approaches. This results in manual, high-effort, and mostly one-off AI audits, necessitating alternative methods. Inspired by other domains such as finance, continuous AI auditing is a promising direction to conduct regular assessments of AI systems. The issue remains, however, since the methods for continuous AI auditing are not mature yet at the moment. To address this gap, we propose the Auditability Method for AI (AuditMAI), which is intended as a blueprint for an infrastructure towards continuous AI auditing. For this purpose, we first clarified the definition of AI auditability based on literature. Secondly, we derived requirements from two industrial use cases for continuous AI auditing tool support. Finally, we developed AuditMAI and discussed its elements as a blueprint for a continuous AI auditability infrastructure.

AuditMAI: Towards An Infrastructure for Continuous AI Auditing

TL;DR

The paper addresses the lack of continuous and integrated AI auditing by proposing AuditMAI as a blueprint for an infrastructure enabling ongoing auditability. It grounds the framework in an explicit AI auditability definition with four key elements and derives four practical requirements from two industrial use cases. The authors describe a three-view architecture—Knowledge, Process, and Architecture—and outline a prototype AuditBox employing semantic-web techniques to integrate auditable artefacts. By combining knowledge management, systematic auditing processes, and architectural components, AuditMAI aims to reduce manual overhead and support preventive, regular AI risk assessments in real-world settings.

Abstract

Artificial Intelligence (AI) Auditability is a core requirement for achieving responsible AI system design. However, it is not yet a prominent design feature in current applications. Existing AI auditing tools typically lack integration features and remain as isolated approaches. This results in manual, high-effort, and mostly one-off AI audits, necessitating alternative methods. Inspired by other domains such as finance, continuous AI auditing is a promising direction to conduct regular assessments of AI systems. The issue remains, however, since the methods for continuous AI auditing are not mature yet at the moment. To address this gap, we propose the Auditability Method for AI (AuditMAI), which is intended as a blueprint for an infrastructure towards continuous AI auditing. For this purpose, we first clarified the definition of AI auditability based on literature. Secondly, we derived requirements from two industrial use cases for continuous AI auditing tool support. Finally, we developed AuditMAI and discussed its elements as a blueprint for a continuous AI auditability infrastructure.
Paper Structure (10 sections, 3 figures, 1 table)

This paper contains 10 sections, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. AI Auditability Definition
  3. Background: Common Discussion Themes in AI Auditability
  4. Proposed AI Auditability Definition
  5. Use-case based AI Auditability Requirements
  6. Use Case Analysis
  7. Auditability Requirements
  8. AuditMAI Framework
  9. Framework Description.
  10. Conclusion

Figures (3)

  • Figure 1: Auditability Definition Key Elements (E1-E4) and Challenges (C1-C2b)
  • Figure 2: A schematic overview of the UCs and derived requirements (R1-R4)
  • Figure 3: AuditMAI: Combining knowledge management, process understanding and architectural design for AI Auditability Infrastructure