Episode #258. The Most Dangerous Thing About AI? Your Permissions, with Stephen Clark, Director of Security at Children's Health in Dallas
Apple Podcasts 
Spotify 
Castro 
RSS Feed An AI agent inherits whatever access the person prompting it already has.
In this episode, Stephen Clark, Director of Security at Children's Health in Dallas, talks about building AI governance inside a pediatric health system where access management, data privacy, and security all sit on his desk. He walks through the three initiatives his team has run for five years: defending the network's edge and interior, protecting data across hundreds of applications and workloads, and putting guardrails around AI so that agents are treated as personified identities rather than unmanaged tools. He explains why Children's Health consolidated on Microsoft Copilot under an E5 license instead of opening the door to every model users were asking for, and what already having PHI classified bought them when they got there. He also makes the case for back-office AI in accounting, finance, and HR, where invoice processing drops from weeks to seconds and a CFO finally sees cash position in real time.
Tune in to hear how a children's hospital adopted AI from a greenfield start and came through it without a recordable incident!
Resources:
Connect with and follow Stephen Clark on LinkedIn.
Follow Children's Health in Dallas on LinkedIn and visit their website!
For more information and ways to increase risk awareness and safety, visit us at www.censinet.com.
Music by David Cosgrove an accomplished composer, musician, producer, and engineer. Listen to his latest project Del Piombo.