PUBLICATION
Addressing AI Trustworthiness in the Architectural Design of Web Systems
Type
Chapter
Year
2026
Authors
Abubaker Gaber
Dr.-Ing. Sebastian Heil
Prof. Dr.-Ing. Martin Gaedke
Research Area
Published in
Distributed AI in the Modern World, 1e: Technical and Social Aspects of Interacting Intelligent Agents
ISBN/ISSN
9780443446801
Download
TBA
Abstract
The increasing integration of artificial intelligence (AI) into web-based systems introduces critical challenges related to trust, transparency, and accountability. Existing AI regulatory frameworks, such as the EU AI Act and NIST AI Risk Management Framework (AI RMF), emphasize the need for systematic approaches to AI trustworthiness, yet practical methods for embedding these principles at the architectural level remain underdeveloped. This chapter envisions a modeling approach based on the WebComposition Architecture Model (WAM) to facilitate trustworthy AI integration into web system at design level. By extending WAM with AI-trustworthy-specific aspects, our approach supports a structured methodology for representing AI services, datasets, and trust relationships within web system architectures. Through this modeling approach, our vision is to enable automated AI risk assessment, compliance verification, and transparency for AI-driven web applications. Two use cases scenarios illustrate the applicability of our WAM extension in industry contexts: (1) AI-as-a-Service (AIaaS), where organisations rely on third-party AI models, and (2) in-house AI deployment, where the entire AI lifecycle is within the organisation, developing and operating the web system. This work contributes to the creation of trustworthy, transparent, and verifiable AI-driven web systems, providing a practical bridge between AI regulatory frameworks and real-world system architecture modeling.
Reference
TBA
