ONE Summit 2024

Leveraging Intel TDX to protect AI workloads, encompassing AI models, parameters, and customer data, from three primary entities—vendors, cloud service providers, and prompt requesters—is a critical aspect of implementing a Zero Trust Architecture for AI Workloads on the Edge. Zero Trust Architecture entails the concept of validating every user and device attempting to access resources on a network, irrespective of their location. When applied to AI workloads on the Edge, Intel TDX serves as a foundational technology to ensure robust security measures. It can safeguard sensitive components such as AI models, parameters, and customer data from potential threats or unauthorized access by any involved party. By employing Intel TDX, which provides hardware-based isolation and security features, it becomes feasible to establish an environment where trust is not assumed, and every access attempt is verified and authenticated. This helps in securing the AI workload against vulnerabilities or breaches that could compromise critical data or models.