Adversarial Attacks and Backdoor Exploitation in Large Language Models: Detection, Forensic Analysis, and Defense Mechanisms

View Sample PDF

Author(s): Rocky Kumar (Poornima University, Jaipur, India)and Joe Arun Raja (Presidency University, Bengaluru, India)
Copyright: 2026
Pages: 38
Source title: Examining Vulnerabilities and Adversarial Exploitation of AI and LLMs
Source Author(s)/Editor(s): Puya Pakshad (Illinois Institute of Technology, USA)and Marwan Omar (Illinois Institute of Technology, USA)
DOI: 10.4018/979-8-3373-8252-4.ch003

Keywords: Artificial Intelligence / Engineering Science Reference / Information Security & Privacy / Security & Forensics

Purchase

View Adversarial Attacks and Backdoor Exploitation in Large Language Models: Detection, Forensic Analysis, and Defense Mechanisms on the publisher's website for pricing and purchasing information.

Abstract

Large Language Models (LLMs) have quickly become cornerstone elements in intelligent systems nowadays, making decisions, automating processes, and doing security operations and interactive applications in a variety of environments. However, their increased integration into critical infrastructures has led to increased concerns regarding malfeasance exploitation by an adversary and/or hidden vulnerabilities. Attackers can exploit these models with prompt-based attacks, backdoors, data poisoning and output manipulation to gain unpermitted access to the model, spread false information, bypass safety filters and to misclassify. These adversarial ways pose a great challenge to the reliability, interpretability and trust degrees when it comes to the AI-driven platforms. This chapter is a detailed look of the adversarial attack surfaces and backdoor exploitation techniques against LLMs. I

The IRMA Community

Research IRM

Adversarial Attacks and Backdoor Exploitation in Large Language Models: Detection, Forensic Analysis, and Defense Mechanisms

Purchase

Abstract

Related Content

IRMA Sponsors