Securing the Invisible Hands: Policy and Guardrails for AI-Driven Infra...
Talkat Open Source Summit Japan + AI_dev: Open Source GenAI & ML Summit + Automotive Linux Summit 2025 · 2025-12-08..12-10·
summary
Claims that Crossplane 2.0 detects config drift within seconds and can automatically remediate, tying AI agent requests into a Kubernetes-style control plane. Demo shows Kyverno rejecting deployments missing resource limits and blocking privileged containers, while composition functions mutate or validate Crossplane composites before provisioning. Runtime detection uses Falco for syscall-level alerts, GitOps (Argo CD) preserves commit history, and a service mesh or network policies isolate offending workloads; the stack is shown working in a prerecorded demo and recommended as a multi-layer practical pattern.
AI-written summary · about
chapters (10)
- 0:00AI-driven infra security intro
- 1:46AI infra risks and challenges
- 4:13Automated guardrails for AI ops
- 8:43Policy as code with Crossplane
- 14:21Real-time monitoring strategies
- 18:35Security tactics for AI-native infra
- 29:29KubeVern policy enforcement demo
- 31:55Security layers and tooling summary
- 34:47AI security resources
- 36:33Q&A on policy enforcement
related talks
13mKeynote: Crossplane Is the Answer! but What Is the Question?
37mPlatformOps with Crossplane
41mAvoiding IAC Potholes with Policy + Cloud Controllers
32mCrossplane: Compose Infrastructure and Services Into...
36mHow to Safely Deploy AI Agents that Write and Manage your IaC
25mDe-risking Deployments in an AI World
25mShift Left in Action: Automating CIS Compliance With Kyverno and CNCF Power Too...
33mUnifying Developer Experience Across Virtual and Cloud Native Worlds