We hear from AI Professor Andy Pardoe about how AI is affecting the cyber security industry, and what to expect from a future dominated by machines
Recently, we were lucky enough to welcome Professor Andy Pardoe to NETbuilder’s London office. He gave an intimate talk about AI in cyber security, which is a major concern not just for everyone in the industry, but ultimately for the end users affected by the threats it poses.
Professor Pardoe began by asking a question: Is AI a friend, a foe, or a frenemy?
“Any truly transformational and disruptive technology can be used for good and bad – and I think this is where we are with AI in cyber security,” he said.
“Where previously it's been very rules-based and reactive, moving forward, AI can really help you get ahead of potential issues and spot these things much sooner.
“That’s the good news.
“The bad news is that the same technologies can make it easy for cyber criminals to attack companies and infrastructure at scale. And that's where things get really tricky.”
Professor Pardoe continued, explaining how AI enables bad actors to attack infrastructure at a scale that has simply not been possible in the past – and it doesn’t stop there.
From large language models (LLMs) contributing to highly personalized social engineering attacks, to AI and the internet of things (IoT) causing serious security concerns for computing at the edge.
Then there’s AI and quantum computing coming together to create a very powerful mechanism.
“One of the fundamental issues we have around AI at the moment is that it can be used for good, but it also can be used for bad. And cyber is not the only example of this,” said Professor Pardoe.
He highlighted the fact that the AI industry has always been very open and willing to share, which means many models have been open source.
“This is great for research and being able to adopt the technology, but it also makes it really difficult to control," he continued.
“How we decide to use this technology – both as an industry, and as a humanity – will be very important as we move forward.”
So how can we defend against AI threats? Professor Pardoe believes the only way is to fight fire with fire.
In the past there have been efforts to press pause on the development of AI to allow time for planning and regulating how we use it. However, its commercial value has ultimately proved more powerful, paving the way for development to continue at this alarming rate.
“It’s terrible because you find yourself in a sort of arms race to get the most advanced technology,” he explained.
“You're going to have AI fighting AI and I think it’s going to be this virtual battle between good AI and bad AI, based on who’s using the technology. The better technology you’re armed with, the better chance you have in defending yourself against this sort of malicious attack.”
So, how can we defend ourselves against these supercharged AI cyber threats? Professor Pardoe thinks the answer lies in agentic AI.
“In the future there will be AI agents that are working almost like an immune system to protect your infrastructure, to protect your company.
“I think that's going to be a new wave of innovation in this space, for sure,” he predicts.
So, what’s the answer? How can we better prepare ourselves for this AI-dominated future? According to Professor Pardoe, there is certainly hope and it all rests on education.
“Like any very transformational technology, it's not the technology that's the problem, it's the people that are using it,” he explained.
“We need to improve the training and education around AI, not just in cyber, but more broadly. Typically, the best outcomes are when we have cross-disciplined teams working on solutions, and I think that's going to be key.”
But while educating good actors to be the best they can be at using AI is essential, the more pressing issue is preventing the bad actors using it to nefarious means.
“Ultimately, we need to figure out how we govern the technology. And while you could argue that the cat is out of the bag a little bit, I still believe there's ways we can build capabilities into the lowest levels of AI infrastructure that will allow us to limit the use of AI for bad actors,” he said.
In closing, Professor Pardoe reflected on his earlier question of whether AI in cyber is a friend, a foe, or a frenemy.
"I think it's all three,” he concluded. “And the challenge is going to be, how do we shape things going forward to make sure we place more focus on the good.”
Learn more about Professor Any Pardoe at his website, or browse our content center to explore more news, views and expert insight from the world of cyber security.