Abstract:

Artificial intelligence is increasingly used to support the analysis of digital evidence by organizing large datasets, detecting patterns and identifying information that requires further human assessment. In digital forensics, such support can be valuable, but it also complicates the understanding of the digital chain of custody. It is no longer sufficient to determine whether evidence has been properly secured, preserved and analysed. It is also necessary to know what role the AI system played, whether its output was verified by a human, and whether the entire process can later be reconstructed [8, 6]. The aim of this article is to present a preliminary model of Responsible BPM for AI-Augmented Digital Chain of Custody, which treats the digital chain of custody as a high-risk organisational process. The model consists of five layers: process, decision, responsibility, audit and traceability, and explainability and human oversight. The article is a research-in-progress contribution and provides a starting point for further validation through expert consultation and investigative scenarios.