AI Sprawl: The Hidden Risk Threatening Your Enterprise Now
summary
Unmanaged AI proliferates from web chat to dozens of copilots and embedded vendors, creating blind spots: models can be deprecated in six weeks, cost different models up to 10x more, and APIs offer 99% uptime vs five‑nines resilience needs. Real failures include a coding agent that deleted a production database (backups saved), Anthropic outages after mass migration, and geopolitical provider bans. Practical steps are discover with browser plugins and endpoint agents, secure with MCP/AI gateways, run living governance logs, and optimize by dynamically routing between models.
AI-written summary · about
chapters (9)
- 0:00Introduction to AI risks
- 1:12Governance and visibility gaps
- 3:20Government influence on AI models
- 5:47AI agent safety and control
- 8:57Ownership of AI biases and hallucinations
- 10:00Securing model context protocol
- 10:45Discovering AI in your environment
- 12:08Governance for real-time AI compliance
- 13:13Optimizing AI usage and resilience
related talks
45mDeploy First, Secure Never: AI Between Productivity and Security Chaos
14mBest Practices for Accelerating Your AI Journey
22mFireside Chat - AI Security Vulnerabilities
43mNavigating the AI Frontier: Investing in AI in the Evolving Cyber...
24mAI Governance for developers: A practical framework for deploying autonomous coding agents safely…
24mTechnical Tracks
24mImplementing Robust AI Security for Enterprise Innovation
52mAI Is Just Software, What Could Possibly Go Wrong?