A conversation with Nitel Fleming Shi, Nitel Chief Technology Officer

While artificial intelligence offers new levels of automation and efficiency, it also introduces risks that can’t be ignored.

Smarter Networks

Managing a network is more complex than ever, which is why AI is becoming a critical tool for enterprises. AI-powered platforms will take a proactive approach to network management, pulling data from different systems and vendors, analyzing it in real-time, and identifying potential issues before they cause disruptions.

What does this mean for network architects? Instead of constantly reacting to outages or security breaches, you get an AI-driven system that helps prevent problems before they start.

The Double-Edged Sword of AI in Cybersecurity

While AI is a game-changer in network security, it also presents new challenges. Fleming compares its potential to nuclear technology: powerful, revolutionary, but dangerous if misused.

“AI capability—it’s like the nuclear bomb,” he warns. “If it gets into the wrong hands, it could really be a problem.”

A recent security breach at an AI startup in China exposed millions of sensitive records due to a critical flaw. Compounding the issue, the AI model was easily manipulated to provide guidance on hacking techniques. For companies handling sensitive customer data, security lapses like these aren’t just concerning—they’re deal-breakers.

“This particular vulnerability is damaging to their brand, of course, but it’s also a wake-up call to their security practices,” Fleming explains.

For network architects, this means AI security isn’t just about protecting your own systems—it’s about scrutinizing any AI-driven tools before integrating them into your infrastructure.

Key Security Practices for AI-Powered Networks

Network architects need to strike a balance between leveraging AI’s benefits and avoiding the pitfalls. Here’s where to start:

  • Know What AI You’re Using – Maintain a full inventory of AI-powered tools, including models, data sources, and integrations. If you don’t know what’s in your environment, you can’t secure it.
  • Set AI Usage Policies—Make sure everyone in your organization understands what AI tools can and can’t be used for. Compliance and security risks can arise when employees start using AI solutions without oversight.
  • Perform Regular Security Checks – AI models and systems should be continuously tested for vulnerabilities, just like any other software.
  • Encourage Collaboration Between AI and Security Teams – AI can’t exist in a silo. Security professionals need a seat at the table when AI-driven solutions are deployed.
  • Keep Up with Regulations – Governments are beginning to crack down on AI security and compliance. Understanding these regulations can prevent costly mistakes down the road.

“As a company, when you’re in a business, establish a policy… putting a data classification in place to make sure they know certain data is top secret,” Fleming advises.

Preparing the Next Generation for AI

Fleming doesn’t just think about AI in the context of enterprise security—he also considers its impact on future generations. When advising his kids about AI, he emphasizes the importance of understanding the bigger picture rather than just learning how to code.

“If you want to learn how to code, I think that’s okay, but try to learn game theory, learn how to solve problems, learn how to build things,” he says. “Not just writing the code, but understanding what’s necessary, the logic behind the problem you’re trying to solve.”

In a world where AI and automation are making coding more accessible than ever, critical thinking and problem-solving skills will set future professionals apart. Fleming believes these skills will remain essential for years to come.

He also stresses the importance of ethics in AI. “AI ethics is another area to focus on,” he says. “Learning what’s right, what’s wrong, and how to detect and measure those outcomes will be very important in the future.”

Fleming’s advice for preparing for an AI-driven world is clear: Focus on problem-solving, strategic thinking, and ethical considerations. These skills will remain valuable no matter how technology evolves.

The Future of AI and Network Security

AI-driven automation is about making networks smarter and more adaptable. Agentic AI, which can independently analyze information and make decisions, is pushing cybersecurity toward a more dynamic and intelligent future.

Fleming describes agentic AI as the next step beyond rigid automation, allowing systems to adapt and respond based on evolving data. “Agentic AI is really about automation, autonomous capabilities to make decisions with more complex information,” he explains. “Not just trained to make decisions, but also able to navigate step by step to understand the outcomes and adjust accordingly.”

In cybersecurity, this means AI-powered agents could work together to detect phishing attempts, track unusual login behavior, and flag potential threats by correlating different signals across the attack surface. Instead of relying on pre-set rules, these systems would reason through different scenarios, much like a human security analyst would.

“The reasoning portion will require us to train AI components to really understand why one step leads to another,” Fleming explains. “It’s like having multiple agents monitoring various aspects of security and sharing intelligence in real-time.”

This level of automation is a major leap forward, but Fleming acknowledges that it also comes with risks. AI security must evolve alongside these advancements to prevent adversaries from exploiting new weaknesses.

Final Thoughts

AI isn’t going anywhere, and for enterprise network architects, it’s both an opportunity and a responsibility. While AI-powered tools can enhance security and efficiency, they also introduce new risks that require constant vigilance.

Fleming sums it up: “Agentic AI is really about automation and autonomous capabilities. It’s able to navigate through decisions step by step, learn from outcomes, and adapt to new information. That’s the kind of intelligence that will be very helpful to us.”

The key takeaway? AI should work for you, not against you. With the right strategy, enterprise network architects can use its power while keeping their infrastructure secure.

Sources:
Interview with Nitel, Chief Technology Officer Fleming Shi Feb. 7, 2025
https://www.cybersecuritytribe.com
https://www.wsj.com
https://www.govinfosecurity.com