tl;dr: As AI becomes central to enterprise operations, cybersecurity must evolve to protect massive, sensitive datasets and ensure trust. Dell’s John Roese and John Scimone discuss how organizations must take a holistic approach to risk management to unlock innovation while staying secure.
It’s impossible to have a conversation about technology today without talking about artificial intelligence. A question on many of our minds is: does security enhance AI, or does AI enhance security?
As we navigate the third year of the generative AI cycle, the intersection of security and trust has become one of the most complex areas for organizations to master.
To explore this vital topic during Cybersecurity Awareness Month, I had the pleasure of sitting down with my colleague, John Scimone, president and chief security officer, who leads all things security and resilience at Dell. We discussed the evolving relationship between AI and security, how organizations can rethink their infrastructure for the AI era and why this moment represents a tremendous opportunity for innovation and progress.
The following has been edited for length and readability.
As a practitioner navigating this space, what’s your view on the relevance of security with AI now? How do you see the current state of security with respect to AI in the enterprise adoption cycle?
Scimone: Security is more essential in an AI world for several reasons. First, from a scale perspective, AI consumes, uses and generates massive amounts of data. Hackers go where the data is. So, the pure volume changes the security imperative. Second is the confidentiality of the data. The models, the data they consume and the data they create are increasingly a company’s most critical intellectual property.
We’ve also seen the scourge of ransomware, where attackers cause pain to extort money. Businesses will become incredibly reliant on AI infrastructure and the data within it. From a hacker’s perspective, the availability of these models and data is a prime target for ransomware groups. Finally, AI engines produce better outcomes when the data is clean, well-managed and labeled. This goes hand-in-hand with data security and management. Across this spectrum, security was essential yesterday, and it’s even more essential today.
We’re telling people to rethink their infrastructure for the AI era because assumptions from the pre-AI era no longer apply. What’s your feeling about how people should approach rethinking infrastructure, including security, as they accelerate adoption of AI?
Scimone: This is an inflection point that sparks a strategic pivot to re-examine infrastructure decisions. The security of data is a key driver for the trust needed in the infrastructure where AI will run. Many customers are reexamining their infrastructure and data strategies. What they’re finding is that it’s increasingly strategic to retain control of their data, not necessarily outsource it to third parties.
This is leading to higher adoption of private clouds and the build-out of modern, zero-trust-minded security architectures. It’s a huge opportunity to think foundationally about risk tolerance and modern practices. Architecturally, it makes the most sense to bring AI to the data, often using private instances of open models rather than sending data to an AI cloud provider. Through a mix of AI models and APIs running on-prem, on-device and at the edge, customers can maintain greater control and secure their intellectual property.
You own the application of AI into Dell’s security ecosystem and must scale with a gigantic threat landscape. What’s your current thinking on AI empowering next-gen security and where the most interesting touchpoints are today?
Scimone: The industry has gone through a cycle of denial and fear, but we are now moving beyond acceptance to excitement. AI possesses the attributes that correlate with the largest pain points the security industry has historically faced. One is the extreme shortage of talent. AI is a massive force multiplier for labor and can help close that gap. Another is spotting anomalies and patterns. In a complex enterprise, finding malicious activity is incredibly hard, but these data science advancements are great at it.
The thing that most excites me is the speed dynamic as we move toward agentic and autonomous operations. Historically, cyber has been an asymmetric space. Bad guys have indefinite time to plan attacks, but defenders must respond in minutes. The future with agentic AI holds the potential for both sides to operate at the same speed. This is inspiring because we can start to close the gap on speed, labor and capability. We now have the opportunity to shift the balance of power to defenders over attackers. What we see today excites us, and what we know is coming excites us even more.
How has your approach to managing risk and preparing the security landscape evolved over the last couple of years, and how do you see it going forward?
Scimone: One thing that stands out is taking a more integrated, whole-of-company approach to risk and opportunity management. Historically, security, legal and IT teams had their own objectives, and things could move very slowly. What’s been exciting at Dell is accelerating the speed of decision-making and removing barriers. We are getting to “yes” quickly, but showing how to do it securely, responsibly and ethically, all while weighing business value and operational risk.
We’ve been able to unlock major value for the company and our customers within weeks or months. And we’ve done it with good security outcomes, without sacrificing our security posture.
Any closing words or a final message for the audience?
Scimone: I appreciate the partnership and the approach of treating this as a company issue, not just an AI, tech or security challenge, but a security opportunity. We will only get it right if we take a holistic approach. Security risks are addressable if we buckle down and commit to moving fast. Fail fast but succeed fast.
View our full conversation in my YouTube series, “AI Insights with John Roese” here: Security & Trust in the AI Era
