Cyber risks facing Ireland discussed at Security Conference

Now, virtually all cyberattacks now feature an AI component. That was the sobering message from Dr. Richard Browne, director of the National Cybersecurity Centre this week. He was speaking at a conference in Dublin, which brought together some of Europe's leading experts in cyber resilience and digital security. It comes after a series of extraordinary disclosures from some of the world's biggest AI companies, including OpenAI, Anthropic, and Meta, about their AI systems going rogue, hacking other companies or organizations. Dr. Richard Browne joins us now, and you're very welcome, Richard, to the program this morning.

Good morning, Sarah.

Um, you've said that every cyberattack now has an AI component. So where are the attacks coming from, and do we have any idea how to address them?

So the, the simple answer is yes, we know how to address them, and they're coming from the places they've always come from. Um, so all the attackers that we've, we've been coming used to seeing for years, nation states, criminal actors, hacktivists, are all using AI and chaining different AI tools together because it's easy, simple, and cheap, and it's an easy way of them to gain greater capacity. That, that has two implications in the short term for us. The first one is that we now have a much narrower window to res-respond to incidents. You know, we, we previously mo- we might've had three, four, seven days for the start of an incident, then it becoming a critical issue. Now we may only have minutes, and that's already the case, unfortunately. The second issue is that the capability gap between the good actors, if you like, capable cyber actors, and the less capable is now collapsing. Everybody has capability. That has real implications for defenders.

Mm. But so y- I know y- and you've sort of, you've echoed it here. You had said that on the threat of AI, we're now in a race against time and against the clock to try and deal with the issues before they become crises. So i- if I'm understanding you correctly, we'll say you become aware of an attempted hacking or a hacking that is underway. You only have minutes to respond, or the company only has minutes to respond, or the state agency, we'll say the HSE, only has minutes to respond. Are they ready to do that if it were to happen tomorrow?

So, so for the most part, the basic general cybersec-security controls still work. That's, I think, the key message that came out of the conference this week. Um, we- no one has really seen an incident that could've been, that couldn't have been prevented by the application of the general security controls. So all of the work that we've been doing for the last decade and, and a half, all of the work that's been happening all over the world in this area still pertains. The issue is that if you already have a vulnerability-

Mm

... if you have a weakness, then that is much more readily exploited, found and exploited, actually, in both cases. So for example, we spent a lot of time in a, over a recent weekend dealing with a vulnerability in a, in a widely used software, or sorry, hardware product, and basically we had to go and find all the potential victims in the country and tell them