Large Language Models in Modern Forensic Investigations: Harnessing the Power of Generative Artificial Intelligence in Crime Resolution and Suspect Identification
Αναστάσιος Νικολακόπουλος, Spyridon Evangelatos, Eleni Veroni, Konstantinos Chasapas, Nikolaos Gousetis, Απόστολος Αποστολάρας, Christos D. Nikolopoulos, Thanasis Korakis
Abstract
Large Language Models (LLMs) have recently captured the attention of the scientific c ommunity. S ince t he global launch of LLM-based chatbots in late 2022, the field h as witnessed a rapid increase in interest from researchers, technology providers and citizens alike. With its wide-ranging applicability, Generative Artificial I ntelligence (GenAI) h as t he p otential to impact various aspects of society, from improving communication and accessibility to transforming industries such as healthcare, education and security. More specifically, in the field of Forensic Science, LLMs could offer significant b enefits as sisting Law Enforcement Agencies (LEAs) and Forensic Practitioners in crime investigations. This paper proposes the implementation of a Retrieval Augmented Generation (RAG) LLM, trained with criminology data, to provide swift and actionable insights into specific incidents, thereby enhancing Forensic Data Analysis and facilitating the daily operations of LEAs.