The cyber space usage has seen multiple levels and types of cyber-crime. The role of digital forensics to collect criminal evidence and analyze it so as to find a path to the cyber-criminal is equally challenging. In this regard, AI tools and techniques have provided a robust weapon. But the main issue with using AI mechanisms is the understandability; interpretability, of a result, is quite difficult in a court of law. In legal aspects, the AI lacks the features of explainability. In this chapter, the authors shed light on the techniques of explainable artificial intelligence techniques that provide methods to enhance the degree of explainability in tracing a cyber-crime so that the court of law delivers the verdict with complete justice. In the future, XAI-driven digital forensics will go a long way in supporting cyber forensics, and this chapter is a sincere attempt to propel the stakeholders in this domain to give a thought in the same direction.
TopIntroduction
In 2009, a popular study in U.S. recognised that the digital forensics discipline has “gone through a quick development process” in the 10 years preceding its distribution. In the official report, three notable perspectives were stated as below (Alam, 2024):
- (1)
Absence of a unbiased accreditation method for the cyber forensic experts,
- (2)
Cyber forensic practises are questioned by other scientific procedures as doubtful
In 2020, the UK Legal Capability network collaborated with the Public Police Boss' Gathering and the Relationship of Police and Wrongdoing Chiefs to create the UK Digital Scientific Science Technique — a solid marker that these gatherings view DF as an insightful and measurable cycle within the domain of policing (Cirqueira,2020).
The exploration recognized three center difficulties confronting cyber forensics (i.e., size of information, complexity of the involved tools used) and several “major problems.” Recognizing that proficient and compelling joining of information is a center capability of DF, that's what the creators expressed “executing further developed innovation for search and survey, counting utilization of computerized reasoning (AI) and AI (ML) strategies” should be “prioritized” (Scientific Capability Organization, Changing Forensics, Public Police Bosses' Committee, Relationship of Police and Wrongdoing and recognized AI as a key arising region where policing work together with industry and the scholarly world (Gopinath,2023).
One more report ready by the U.S. President's Gathering of Guides and Innovation (PCAST) investigates issues within present day digital forensics practice. This report tracked down that measurable discipline in the U.S. needed “quality culture” and that essentially all wrongdoing research centers were intently restricted with the arraignment of cases (Alam, 2022). That's what high Court held “the admissibility of logical master declaration relied upon its logical reliability”. The creators additionally pushed that criminological cycles need to face logical examination by being repeatable, reproducible, and precise. In a perfect world, any system ought to be level headed with “exact definitions” of “include recognizable proof, correlation and coordinating”. As evident, this is absurd, and “the black box in the analyst's head can't be analyzed straightforwardly for its central premise in science,” the creators stress the requirement for observational investigation of analysts' exhibition in “black-box studies” to decide a probability of blunder in view of the summed up exhibition of the scientific expert partner (Hall,2022). A really explainable AI with a rationalizable “black-box” could give more noteworthy trust in emotional criminological approaches, by being both straightforward in process and giving legitimate and coherent thinking.