A.I. makes cybercrime faster, not smarter; Donegal firms should tighten oversight

ESET Ireland has warned that the growing use of AI coding assistants by criminals is changing the pace of cyberattacks, making it easier to automate work that previously required more time, skills, and manpower.

The warning follows public reporting this week describing a case in which an attacker used an AI chatbot to support cyberattacks on government systems, including identifying weaknesses and speeding up scripting and automation.

George Foley, spokesperson for ESET Ireland, a cybersecurity company, said the headline detail is not the country involved, but the method.

“This is what’s changing. The grunt work is getting easier to industrialise. If a criminal can use an AI tool to move faster, iterate faster and automate more, the gap between ‘trying it’ and ‘doing damage’ gets smaller,” he cautioned.

Foley said Donegal organisations should not treat this as an “AI panic” story, but they should be prepared.

“AI doesn’t magically break into networks. The usual doors still matter, weak passwords, excessive access, unpatched systems, people clicking what they shouldn’t. AI just helps attackers work through those opportunities at speed.”

ESET research has previously reported on PromptLock, a ransomware variant that uses generative AI as part of its execution flow, as an example of how the misuse of AI is already moving from theory into practice.

READ NEXT: Crana Swimmers generously donate €1,300 to Foyle Hospice

Foley said the shift lands at a time when more Irish organisations are being pushed towards board-level accountability for cybersecurity under the EU’s NIS2 direction of travel, regardless of sector.

“The organisations that will cope best are the ones that have basics nailed down and ownership nailed down. Who can access what? Who approves changes? Who gets alerted when data starts moving in ways it shouldn’t? And who runs the response when something goes wrong?”

He said the priorities for most organisations are straightforward: tighten identity and access; reduce admin privileges; patch known weaknesses quickly; monitor for unusual data movement; and make sure staff know what modern phishing and social engineering looks like in 2026.