Towards Verifiable Traceability in AI Systems: Operationalizing Logging and Auditability under the EU AI Act

This preprint proposes a minimal, technology-neutral model for verifiable traceability in AI systems. It addresses the gap between high-level regulatory obligations under the EU AI Act and their practical implementation by defining auditable events, integrity-preserving event structures, and independently verifiable evidence packages. The objective is to support reconstruction, auditability, integrity validation, and independent verification of AI system behavior without prescribing a specific technology or implementation.