James Nettesheim, Block | Block Builderlab Launch Event @siliconangle
James Nettesheim, Block | Block Builderlab Launch Event  @siliconangle
Uploaded November 2026 | Updated November 2026, 2 days ago
In this interview from Block Builderlab Launch Event 2026 in San Francisco, James Nettesheim, chief information security officer at Block, joins theCUBE's Paul Gillin and Alison Kosik to discuss securing the leap from AI that answers questions to AI that takes action. Nettesheim argues that enterprises don't need to reinvent the wheel because time-tested principles such as least privilege, isolation and egress controls still apply. In Builderlab, agents become full-fledged identities with delegated, just-in-time permissions. They operate inside hardened cloud virtual machines, and every action lands in a cryptographically signed, immutable ledger that teams can replay and review.

Additionally, Nettesheim shares how Block is building Information Flow Control, a future capability in which security policy travels with sensitive data and blocks it from leaking to unauthorized destinations. He details how AI is already reshaping Block's own defenses, from autonomous code reviews to alert triage. Judge agents challenge findings, and each alert gets the undivided attention that a busy human analyst can't always give. For the first time, Nettesheim believes defenders can match attackers' machine speed, tilting a historically asymmetric fight toward the good guys. The discussion also covers third-party agent risk and the need to replace one-time checkbox reviews with real-time monitoring. It also addresses Block's pursuit of ISO 42001 and the rise of deepfake-driven fraudulent hiring schemes that demand defense in depth. From creating paved paths that curb shadow AI to treating trust as a security challenge, Nettesheim offers a practical playbook for enterprises looking to scale agents with confidence.

Find more SiliconANGLE news and analysis siliconangle.com
Follow theCUBE's wall-to-wall event coverage siliconangle.com/events
Learn about the latest theCUBE events thecube.net

00:00 - Intro
00:07 - Enterprise Security for Agentic AI
03:22 - Preventing Leaks With Information Flow Control
05:49 - Security and Accountability for Scaling AI Agents
08:35 - Evaluating AI for Code Reviews and Cybersecurity
11:33 - Third-Party AI Agent Monitoring and Security Standards
13:54 - Defending Against AI Attacks and Fraud
16:39 - Managing Security and Trust in Enterprise AI

#theCUBE #BlockBuilderlab #theCUBEresearch #Block #AI #Cybersecurity
James Nettesheim, Block | Block Builderlab Launch EventRoss Armstrong, Phoenix AI | Cloudera EVOLVE26AI versus AI: The cybersecurity race heats up at Fal.Con 2026Rehan Jalil, Veeam | theCUBE + NYSE Wired: Cyber Security Leaders
SiliconANGLE theCUBE |

James Nettesheim, Block | Block Builderlab Launch Event

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER