Digital Forensics AI: Evaluating, Standardizing and Optimizing Digital Evidence Mining Techniques

Abstract

The impact of AI on numerous sectors of our society and its successes over the years indicate that it can assist in resolving a variety of complex digital forensics investigative problems. Forensics analysis can make use of machine learning models’ pattern detection and recognition capabilities to uncover hidden evidence in digital artifacts that would have been missed if conducted manually. Numerous works have proposed ways for applying AI to digital forensics; nevertheless, scepticism regarding the opacity of AI has impeded the domain’s adequate formalization and standardization. We present three critical instruments necessary for the development of sound machine-driven digital forensics methodologies in this paper. We cover various methods for evaluating, standardizing, and optimizing techniques applicable to artificial intelligence models used in digital forensics. Additionally, we describe several applications of these instruments in digital forensics, emphasizing their strengths and weaknesses that may be critical to the methods’ admissibility in a judicial process.