When AI Agents Cross the Boundary: Securing the AI Factory

About this episode

In July, OpenAI disclosed that models undergoing a cybersecurity evaluation escaped their testing environment, exploited a zero-day vulnerability and ultimately reached Hugging Face’s production infrastructure.

The incident raised a difficult question for every organization deploying autonomous AI: What happens when an agent crosses a boundary — not because it was instructed to attack, but because doing so helps it complete the objective it was given?

Recorded before the incident became public, this conversation from Cisco Live! features WWT's Istvan Berko, Cisco's Vikram Varakantum and NVIDIA's Trevor Caulder, who discuss what it takes to secure AI at machine speed. They examine the security challenges it has now made impossible to ignore, and discuss why enterprises must secure AI from the silicon and network to the workload, model and identity of every agent.

Support for this episode provided by: VAST Data

More about this week's guests:

Istvan Berko leads AI Cyber and Innovation at World Wide Technology, where he works with organizations building secure AI systems for production. With more than 25 years in cybersecurity, risk and governance, he brings a practical perspective on the identity, infrastructure and operational challenges shaping the next generation of AI.

Istvan's top pick: Everyone Sees the Risk. Nobody Owns It.

Vikram Varakantam is Senior Director of Product Management at Cisco, where he leads the strategy behind the company's AI security portfolio. With experience across AI, cloud and application security, he works at the intersection of product, infrastructure and enterprise AI, helping organizations secure AI as it moves into production.

Vikram's top pick: Securing AI When Agents Move Faster Than Humans Can Respond

Trevor Caulder is a Principal Technologist at NVIDIA, where he works with partners across the AI ecosystem to advance the infrastructure behind secure AI. With more than 25 years of experience spanning silicon, networking and enterprise systems, he brings a hardware-first perspective to securing AI at scale.

Trevor's top pick: Breaking Data Silos: How Private Inference Unlocks GPU ROI on Sensitive Data

The AI Proving Ground Podcast leverages the deep AI technical and business expertise from within World Wide Technology's one-of-a-kind AI Proving Ground, which provides unrivaled access to the world's leading AI technologies. This unique lab environment accelerates your ability to learn about, test, train and implement AI solutions. 

Learn more about WWT's AI Proving Ground.

The AI Proving Ground is a composable lab environment that features the latest high-performance infrastructure and reference architectures from the world's leading AI companies, such as NVIDIA, Cisco, Dell, F5, AMD, Intel and others.

Developed within our Advanced Technology Center (ATC), this one-of-a-kind lab environment empowers IT teams to evaluate and test AI infrastructure, software and solutions for efficacy, scalability and flexibility — all under one roof. The AI Proving Ground provides visibility into data flows across the entire development pipeline, enabling more informed decision-making while safeguarding production environments.