Enterprise AI deployment is creating a security blind spot traditional architectures can’t handle

Enterprise AI deployment has fundamentally expanded the cybersecurity attack surface, turning data pipelines, model training environments, identity management systems and supply chains into prime targets that traditional security architectures were never designed to defend.

The shift from conventional applications to AI factory infrastructure introduces a new class of exposure that organizations are only beginning to reckon with — a challenge that will be paramount at this year’s Dell Technologies World. Where a conventional application might have had a single entry point, the AI factory presents a sprawling, interconnected set of vulnerabilities spanning model inferencing and training data to prompt injection and agentic workflows, according to Steve Kenniston (pictured), senior cybersecurity evangelist for portfolio marketing at Dell Technologies Inc.

“AI changes the whole game,” Kenniston said. “There’s the model inferencing. There’s the model training data. There are systems where people can do things like prompt injection. There’s identity management that needs to be thought about. These things are changing so fast. There’s a whole group of things that actually change from an attack surface standpoint that you want to make sure you have locked down as you’re building out this brand new application. Every new application has a new attack surface.”

Kenniston spoke with theCUBE’s Dave Vellante at the Securing the AI Factory with Dell Technologies and Intel event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how the AI factory expands and transforms the enterprise attack surface, why security must be hardwired into systems from day one and how the human element of recovery is too often overlooked. (* Disclosure below.)

Enterprise AI deployment demands design-in security, not bolt-on

The urgency of securing the AI factory is being driven home in real cases of enterprise AI deployment. Dell’s services organization now reports roughly 85% to 90% of AI projects being halted mid-implementation because the security team had not been included from the outset — a sign that security remains an afterthought, according to Kenniston.

“What it’s saying is that it is still being bolted on,” he said. “The last thing you want to do is get to the five-yard line and have someone from security go, ‘Stop, stop, stop. We haven’t vetted this. We haven’t looked through this. We don’t understand what’s going on.’ You want to make sure that [security is] a part of it.”

Dell’s approach treats the AI factory as a single integrated security surface rather than a collection of individual components — validated across compute, networking, storage and data layers to close the gaps where breaches typically occur. That same thinking extends to the supply chain, with customers now raising integrity questions in nearly every briefing, compared with just one in 20 two years ago, Kenniston noted. Zero-trust rounds out the strategy, moving security teams beyond firewall thinking toward granular identity controls: who has access, when and from where.

“At Dell, we integrate security into everything that we do, right from the supply chain, through the chips, right to the device that gets delivered to you,” Kenniston said. “Where [it] breaks down in a system like this is where security falls between the cracks.”

Stay tuned for the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the “Securing the AI Factory with Dell Technologies and Intel” event.

(* Disclosure: TheCUBE is a paid media partner for the “Securing the AI Factory with Dell Technologies and Intel” event. Neither Dell, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE