Chain of Custody and Forensic Admissibility of Blockchain-Logged Machine Learning Detections in Smart Meter Cyber-Crime Prosecution

• Adaan Ahsun

  Covenant University

  Author

The integration of machine learning (ML) for detecting false data injection attacks in smart meters has demonstrated considerable technical promise, yet the prosecutorial viability of such detections remains underexplored. This study addresses the critical gap between algorithmic detection and courtroom admissibility by developing and evaluating a blockchain-anchored chain of custody framework for ML-generated evidence in smart meter cyber-crime prosecutions. The research employs a design-based methodology combining retrospective analysis of 2,847 simulated attack events with prospective validation of a prototype forensic logging architecture. Key findings indicate that the proposed framework achieved 89.4% accuracy in maintaining evidentiary integrity across simulated legal challenges, with a 94.2% success rate in establishing unbroken chain of custody documentation. The blockchain layer demonstrated 99.7% reliability in tamper detection, while ML detection components maintained 91.3% precision and 87.8% recall under adversarial conditions designed to mirror real-world attack sophistication. The study concludes that integrating cryptographic audit trails with ML detection outputs significantly enhances prosecutorial readiness, though interpretability constraints of complex models require human expert validation protocols. The framework offers practitioners a replicable architecture for generating legally defensible evidence from automated detection systems, while policymakers gain empirical foundations for standards governing algorithmic evidence in critical infrastructure protection.

• pdf

Copyright (c) 2026 Adaan Ahsun (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.