Kubernetes MLSec - Securing AI in Space

Explore the critical intersection of AI security and cloud native technologies in this 35-minute conference talk by Francesco Beltramini and James Callaghan from ControlPlane. Delve into the challenges of securing AI in space, focusing on the vulnerabilities in operational and data security within AI technologies. Examine the risks associated with feeding sensitive data into models running on external systems, and learn about high-profile breaches involving cross-customer data leaks. Discover how cloud native solutions can address these concerns through a comprehensive exploration of Kubernetes-powered MLOps threat modeling. Witness a demonstration of breaking into and poisoning a Kubernetes model-training environment, and understand the inherent dangers of feeding data into Large Language Models (LLMs) and training ML models. Gain valuable insights into cloud native architectural and procedural remediation strategies to enhance AI security in space-related applications.