Compliance of AI Systems
Julius Schöning, Niklas Kruse
TL;DR
The paper addresses the challenge of ensuring AI systems comply with the forthcoming EU AI Act, with a focus on data-set legality and edge/deployed AI. It proposes a platform-based framework that integrates Explainable AI (XAI) with regulatory requirements to enable early legal assessment and ongoing trustworthiness verification. It contributes by categorizing XAI across Ex-Ante, Ex-Nunc, and Ex-Post stages, detailing a two-part platform design with automatic data-set legality checks, and illustrating high-risk considerations through an industrial example. The approach aims to reduce redevelopment costs, improve transparency and imputability, and support responsible, regulatorily aligned deployment of embedded AI.
Abstract
The increasing integration of artificial intelligence (AI) systems in various fields requires solid concepts to ensure compliance with upcoming legislation. This paper systematically examines the compliance of AI systems with relevant legislation, focusing on the EU's AI Act and the compliance of data sets. The analysis highlighted many challenges associated with edge devices, which are increasingly being used to deploy AI applications closer and closer to the data sources. Such devices often face unique issues due to their decentralized nature and limited computing resources for implementing sophisticated compliance mechanisms. By analyzing AI implementations, the paper identifies challenges and proposes the first best practices for legal compliance when developing, deploying, and running AI. The importance of data set compliance is highlighted as a cornerstone for ensuring the trustworthiness, transparency, and explainability of AI systems, which must be aligned with ethical standards set forth in regulatory frameworks such as the AI Act. The insights gained should contribute to the ongoing discourse on the responsible development and deployment of embedded AI systems.