Enhancing Trust and Transparency in Digital Forensics and Incident Response using Explainable AI (XAI): An Extensive Survey

作者：Pallavi Shingade, Prashant Dhotre, Prashant Kharat · 发表于：2025 IEEE International Conference on Blockchain and Distributed Systems Security (ICBDS) · 年份：2025 · DOI：10.1109/icbds67396.2025.11377825

The efficiency of detecting, analyzing and mitigating cyber incidents has been enhanced by increasing use of Artificial Intelligence (AI) in Digital Forensic and Incident Response (DFIR). Growing dependence on complex machine learning models raises critical concern about trust, transparency and legal admissibility as their internal working often remain opaque. Across forensic and legal domains accuracy alone isn't enough because results must be reliable, interpretable and defensible to meet requirements of analyst, legal experts and Courts. Explainable AI (XAI)plays vital role in building trust by making AI decisions transparent, verifiable and secure from tampering. This survey focuses on improving trust and accountability in DFIR with XAI. It reviews widely used methods as SHAP, LIME and interpretable rule-based models used in multiple matured domains. This work also evaluates advantages and limitations of current approaches in this area. Furthermore, it also focuses on metrics that assess technical performance and forensic applicability. Finally, survey identifies several research gaps including domain standards, secure logging & robustness against any kind of manipulation. Survey outlines potential research areas to develop trustworthy and legally admissible AI driven framework.