Deploy First, Secure Never: AI Between Productivity and Security Chaos
summary
Overview of security risks introduced by deploying AI across companies, covering LLMs, autonomous agents, MCP (Model Context Protocol) and multi-agent chains. Presents real demonstrations of vulnerabilities, maps risks using OWASP’s Top 10 for LLM Applications and the 2026 Top 10 for Agentic Applications, and offers concrete mitigation suggestions plus a compliance focus on implications of the EU AI Act for deployments.
AI-written summary · about
chapters (10)
- 0:00AI security risks
- 3:32Neobank architecture attacks
- 5:45Direct prompt injection demo
- 12:54Indirect prompt injection demo
- 18:39Agent workflow manipulation
- 27:05AI supply chain attacks
- 33:47Uncontrolled consumption fixes
- 39:02Remediations and best practices
- 40:27EU AI regulation
- 42:10Q&A
related talks
14mAI Sprawl: The Hidden Risk Threatening Your Enterprise Now
33mUnlocking Agentic AI: [...]
08mWhy AI Governance Matters More Than Ever
37mFundamentals of AI safety and security
22mFireside Chat - AI Security Vulnerabilities
44mWhat Security Teams Need to Know and Do
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
57mAI Applications Security Puzzle