Zero Trust for AI: Enabling Secure Workload Execution and Data Protection

Abstract

This work seeks to determine how integrating Zero Trust Architecture (ZTA) with artificial intelligence (AI) systems in retail can help strengthen data security, keep up with regulations and execute tasks more securely. It looks at the risks in using AI, for example, poisoning of AI models, data being leaked and adversarial attacks and suggests a method focused on checking each user’s identity, providing them with access only when needed and continuously checking the system. Research has revealed that adopting Zero Trust security lowers risks, adds trust to AI systems, and supports secure operations for customers. This paper advises implementing strong governance, targeted AI-based security and following compliance standards for a secure and ongoing switch to digital in the retail industry.