A new arXiv systematic review examines whether explainable AI methods can satisfy the EU Right to Explanation in practice. The paper focuses on GDPR Article 15(1)(h), Article 86 of the EU AI Act, and related legal instruments.
The authors started from 2,643 records and reviewed 57 full texts. Only 19 papers showed substantive integration of both legal and technical perspectives, suggesting that much of the field still treats law and explainability as separate conversations.
The review identifies recurring problems, including misidentification of the GDPR legal basis and limited engagement with newer legal developments. It also emphasizes the difference between the form of an explanation and the technical process that produced a decision.
The practical implication is that explainability tools cannot be judged only by whether they generate a plausible rationale. They must map to legal rights, affected users, and contestable decisions.