Skip to main content

패치 전에 두 가지 중요한 PHP 결함이 누출되었습니다. 조직이 지금 해야 할 일.

What does it take to deploy 200 AI agents safely?

2026년 11월 19일 · 23:17 min

Featured guests

Mani Masood

Mani Masood

VP of Information Security and CISO, AMN Healthcare

Hosts

Johnny Hand

Johnny Hand

Global CISO · TrendAI™

Dustin Childs

Dustin Childs

Head of Threat Awareness · ZDI

Available wherever you get your podcasts

TrendAI Security Brief Episode 16

About the episode

What happens when your organization sets a target of 200 AI agents, not just built, but deployed and safely sustained? Mani Masood, VP of Information Security and CISO at AMN Healthcare, faced that mandate. His governance approach held up for the first dozen agents and then began to stretch as deployment scaled. In conversation with hosts Johnny Hand and Dustin Childs, Mani shares hard-won lessons on effective AI governance and explains why passing an initial review doesn't guarantee an agent will keep behaving as expected. He shares what he would do differently if his agent target grew fivefold tomorrow, and why AI governance should start with what your cloud platforms can enforce.

What you'll learn
  • How to structure governance around a business risk profile instead of a control profile
  • Why risk ownership for high-risk AI projects must be shared across executives, development, security, and legal
  • Mani's prediction for the biggest AI security surprise of 2027

Episode chapters

  • 01:41 — Setting an aspirational target of 200 AI agents
  • 03:33 — Lessons from the dot-com and cloud eras
  • 07:51 — Unbudgeted risks: stolen API keys and insider threat
  • 12:25 — Scaling agents: start with the platform, not the policy
  • 15:51 — Why AI risk ownership must be shared
  • 19:24 — The biggest AI security surprise of 2027

"If we want to experiment with a high-risk technology, then that risk has to be shared."

— Mani Masood, VP of Information Security and CISO, AMN Healthcare

Hosts

Johnny Hand

As the Global CISO for TrendAI™ and veteran of enterprise security, Johnny brings a security leader’s lens to every conversation, separating exploitable risk from theoretical noise. Johnny is the person you want explaining AI-era threats before your board asks about them.

Johnny Hand
Dustin Childs

Dustin Childs

As Head of Threat Awareness for the TrendAI™ Zero Day Initiative (ZDI), Dustin brings a vulnerability researcher's eye to the AI security conversation. He’s dedicated to making sure the industry understands exploit development before it becomes a crisis.

More episodes

  • How close are we to an autonomous SOC?

    Episode 15 · 2026년 11월 5일

    How close are we to an autonomous SOC?

    Jon Oltsik explains why agentic SOCs are inevitable, why full autonomy is hype, and how to pilot AI in security operations with human judgment in the loop.

  • How do you plan for AI system failure?

    Episode 14 · 2026년 10월 20일

    How do you plan for AI system failure?

    What happens when AI stops working? Maria Haight explains how human know-how, clear ownership and data backups keep businesses running.

  • AI Security Brief Episode 13 with guest Tyler Shields

    Episode 13 · 2026년 10월 8일

    Is trust holding back automated remediation?

    Tyler Shields, CMO at Allstacks and a veteran cybersecurity industry analyst, unpacks the “AI trust paradox” and shares a practical path forward.

  • AI Security Brief Episode 12 with guest Jason Cradit

    Episode 12 · 2026년 9월 24일

    Can you trust AI with critical infrastructure?

    Jason Cradit has spent more than 25 years working across energy, pipelines, aerospace, and defense.

  • AI Security Brief Episode 11 with guest Morgan Adamski

    Episode 11 · 2026년 9월 10일

    Is a closed or an open AI model more trustworthy?

    Morgan Adamski works directly with boards, CISOs, and security teams navigating AI adoption.

  • AI Security Brief Episode 10 with guest Zach Evans

    Episode 10 · 2026년 8월 27일

    What does hospital downtime teach us about building AI-native organizations?

    Is your organization truly AI native, or did you just bolt AI onto existing infrastructure? Your answer could be an indicator to how much risk you’re carrying without realizing it.