Manage episode 492724345 series 3604885
This podcast explores MITRE's SAFE-AI framework, a comprehensive guide for securing AI-enabled systems, developed by authors such as J. Kressel and R. Perrella. It builds upon established NIST standards and the MITRE Adversarial Threat Landscape for Artificial Intelligence Systems (ATLAS)™ framework, emphasizing the thorough evaluation of risks introduced by AI technologies. The need for SAFE-AI arises from AI's inherent dependency on data and learning processes, contributing to an expanded attack surface through issues like adversarial inputs, poisoning, exploiting automated decision-making, and supply chain vulnerabilities. By systematically identifying and addressing AI-specific threats and concerns across Environment, AI Platform, AI Model, and AI Data elements, SAFE-AI strengthens security control selection and assessment processes to ensure trustworthy AI-enabled systems.
Sponsors:
https://airiskassess.com
198 episodes