Security Challenges and Mitigation Strategies in Generative AI Systems
Year: 2025
Type: article
Abstract: This article examines the critical security challenges and mitigation strategies in generative AI systems. The article explores how these systems have transformed various sectors, particularly in financial markets and critical infrastructure, while introducing significant security concerns. The article analyzes various types of adversarial attacks, including input perturbation and backdoor attacks, and their impact on AI model performance. Additionally, it investigates model stealing threats and data privacy concerns in AI deployments. The article presents comprehensive mitigation strategies, including advanced defense mechanisms, enhanced protection frameworks, and secure access control implementations. The article findings demonstrate the effectiveness of integrated security approaches in protecting AI systems while maintaining operational efficiency. This article contributes to the growing body of knowledge on AI security by providing evidence-based strategies for protecting generative AI systems across different application domains.
Author: Satya Naga Mallika Pothukuchi
Institution: University College for Women
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4408247299