Authors: Wickramasekara, Akila and Scanlon, Mark

Publication Date: April 2024

Publication Name: Proceedings of the 12th International Symposium on Digital Forensics and Security

Abstract:

The increasing frequency and rapidity of criminal activities require faster digital forensic (DF) investigations. Currently, most DF phases involve manual procedures, requiring significant human effort and time, often facing evolving requirements. This paper proposes an integrated framework employing AutoGen Artificial Intelligence (AI) agents and Large Language Models (LLMs) such as LLAMA, and StarCoder. The suggested framework utilizes AI agents and LLMs to perform tasks articulated in natural language by a human agent. The proposed architecture presents a significant advantage by alleviating the investigative workload and shortening the learning curve for investigators. However, it is still combined with risks such as information accuracy, hallucination impact, and legal barriers. Although, this research contributes to the ongoing discourse on optimizing DF processes in response to the evolving landscape of criminal activities and the corresponding demands placed on investigative resources.

Download:

BibTeX Entry:

@inproceedings{wickramasekara2024DFAutoGenAI,
author={Wickramasekara, Akila and Scanlon, Mark},
title="{A Framework for Integrated Digital Forensic Investigation Employing AutoGen AI Agents}",
booktitle="{Proceedings of the 12th International Symposium on Digital Forensics and Security}",
year=2024,
pages = {},
month=04,
publisher={IEEE},
abstract={The increasing frequency and rapidity of criminal activities require faster digital forensic (DF) investigations. Currently, most DF phases involve manual procedures, requiring significant human effort and time, often facing evolving requirements. This paper proposes an integrated framework employing AutoGen Artificial Intelligence (AI) agents and Large Language Models (LLMs) such as LLAMA, and StarCoder. The suggested framework utilizes AI agents and LLMs to perform tasks articulated in natural language by a human agent. The proposed architecture presents a significant advantage by alleviating the investigative workload and shortening the learning curve for investigators. However, it is still combined with risks such as information accuracy, hallucination impact, and legal barriers. Although, this research contributes to the ongoing discourse on optimizing DF processes in response to the evolving landscape of criminal activities and the corresponding demands placed on investigative resources.}
}

---