How should AI decisions be explained? Requirements for Explanations from the Perspective of European Law
It addresses the problem of aligning AI explanations with legal standards for stakeholders like policymakers and developers, but it is incremental as it builds on existing XAI-taxonomies and legal frameworks.
This paper investigates the requirements for explainable AI (XAI) methods from the perspective of European law, including GDPR and product safety regulations, and concludes that current XAI methods do not fully satisfy these requirements, particularly in terms of correctness and confidence estimates.
This paper investigates the relationship between law and eXplainable Artificial Intelligence (XAI). While there is much discussion about the AI Act, for which the trilogue of the European Parliament, Council and Commission recently concluded, other areas of law seem underexplored. This paper focuses on European (and in part German) law, although with international concepts and regulations such as fiduciary plausibility checks, the General Data Protection Regulation (GDPR), and product safety and liability. Based on XAI-taxonomies, requirements for XAI-methods are derived from each of the legal bases, resulting in the conclusion that each legal basis requires different XAI properties and that the current state of the art does not fulfill these to full satisfaction, especially regarding the correctness (sometimes called fidelity) and confidence estimates of XAI-methods. Published in the Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society https://doi.org/10.1609/aies.v7i1.31648 .