How Microsoft Secured AI Agents: From Banning OpenClaw to Safe Enablement
a16z Podcast · rss · 2026-08-21
a16z's Joel De La Garza talks with Aaron Zollman, Deputy CISO at Microsoft Gaming, about how security teams can embrace AI agents without losing control. He walks through Microsoft's journey with OpenClaw: the initial instinct to ban it, then figuring out how to make it safe to use.
The discussion covers what agents mean for identity, permissions, containerization, and monitoring, and how AI shifts the CISO's role from saying "no" to safely enabling new technology. They also explore whether AI could help defenders patch vulnerabilities as fast as they're discovered, and why new AI threats don't make old security problems go away.
More from Companies & People
- Critique: Anthropic's Fear Mongering and Rise of the Machine Narratives — gerardsans · 2026-08-21
- Big Tech Meeting: Everyone Is Using Claude for Work — teodorio · 2026-08-21
- Anthropic to let enterprises self-host 30-day logs, retains no copy — rohanpaul_ai · 2026-08-21
- Real Brokerage rolls out Leo 2.0 to automate 80% of real estate tasks — brianrkelly · 2026-08-21
- Can Chinese LLMs Offer 100T Tokens for Free? Cost Advantages Spark Speculation — teortaxesTex · 2026-08-21
- Spent 3 Hours Debugging a Failed AI Model: The Problem Was My Own Assumptions — ingliguori · 2026-08-21