ETSI Releases Security Specification for AI Computing Platforms

• •ETSI published TS 104 033 to define security requirements for AI computing platforms.
• •The specification addresses security functions and components to protect AI models and data.
• •It aims to support compliance with the EU AI Act and mitigate threats like data leakage.

The European telecoms standards body ETSI has published TS 104 033, defining security requirements for AI computing platforms. This specification aims to establish a systematic security baseline for platforms hosting AI applications in data centers and edge computing environments. It addresses security functions, platform components, and interfaces designed to protect AI models, datasets, training processes, and inference workloads. The specification also outlines mandatory security requirements to mitigate threats like model extraction and data leakage. ETSI emphasizes the importance of securing the computing platform as AI systems are increasingly deployed across various use cases. The release is intended to support compliance with regulatory requirements, including the EU AI Act. This development builds on previous frameworks established by ETSI for AI computing platform security.