In this episode of Zero Downtime, Joye Purser sits down with Kristen Beneduce, former Deputy CISO at Nextdoor and incident response lead at Sandia National Labs, to talk about what it actually takes to secure AI systems before they become a liability. Kristen brings a rare perspective, having worked both inside a national lab with a deeply embedded security culture and in the private sector where security still has to fight for its seat at the table. She breaks down why AI agents need to be treated like users, complete with access controls, monitoring, and accountability, and why the organizations getting this right are building faster, not slower.

They cover:

• Why security controls make businesses more efficient, not less
• How to build tighter linkages between security and engineering teams
• What good AI incident re ...