How to Secure the AI Stack from Code to Runtime
summary
Explains a concrete Sonar + Wiz integration that ingests Sonar static-analysis findings into the Wiz security graph to map where a repo issue becomes a runtime risk, enabling code-to-cloud impact and cloud-to-code tracebacks so teams can assign PRs and fix fast. Claims AI can turn zero-day exploits from weeks into days or hours, describes building three agents (detection, risk triage, remediation) and a Win MCP program to automate agent-to-agent workflows for faster, deterministic guardrails.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 0:51How AI-Driven Development Is Reshaping Security
- 1:34AI Accelerates Delivery but Increases Risk
- 2:42New AI Security Risks: Visibility and Scale Challenges
- 3:40Attackers Weaponizing AI for Faster Cyberattacks
- 5:12Expanding Security Ownership Across Teams
- 6:14Security as a Team Sport in DevSecOps
- 8:22Shifting Security Reviews Earlier in the SDLC
- 9:04Security Guardrails in CI/CD and AI Coding Tools
- 10:14Sonar and Wiz: Connecting Code Security to Cloud Runtime
- 11:24Unified Security Visibility Across Code and Cloud
- 12:28Security Graph: Mapping Code Vulnerabilities to Cloud Risk
- 13:19Tracing Cloud Security Risks Back to Developers
- 14:27Deterministic vs Probabilistic Security Approaches
- 16:42Why Risk Context Matters More Than Surface Findings
- 19:08The Future: Agentic AI Security Workflows
- 20:43Winning with AI Means Using AI Defensively
- 21:19AI Agents for Vulnerability Triage and Remediation
- 22:18Next Step: Agent-to-Agent Security Collaboration
- 23:01Sonar and Wiz Roadmap for AI Security
- 24:06Closing Thoughts on AI Security and Partnership
related talks
19mShielding innovation: Safeguarding cloud and AI development
28mMaximizing The Impact Of AppSec with Shift Left And AI
28mAI for Everything After Code
58mSecuring Code in the Age of AI
18mCommit Secure AI: Embedding Security in Code, Pipelines, and Production
42mAI for security - friend or foe?
28mBuilding cloud security programs
19mKeynote: Towards Building Safe and Secure Agentic AI