More than half of Irish business leaders worried over 'sophisticated cyber attacks' by AI

Business leaders across Ireland have been developing increasingly unfavourable views on the role of artificial intelligence (AI) in society amid concerns over “sophisticated cyber attacks”, a study has indicated.

One in four Irish business leaders “now see AI study a business risk, rather than a benefit”, the monthly AI Pulse survey conducted in September by technology consultants Expleo. It represented a five percentage point increase on August’s findings.

A further 56 per cent “are concerned the cybersecurity risks posed by AI”, while 29 per cent “say their inability to successfully address cybersecurity and data protection risks has negatively impacted their opinion of AI”.

The tech advisory firm measured sentiment on AI use and implementation from 200 businesspeople across Ireland, Britain, France and Germany.

The study coincided with Cybersecurity Awareness Month and follows a spate of high-profile instances that have led to safety concerns over the proliferation of AI.

OpenAI , which runs the ChatGPT chatbot service, has disclosed incidents over recent weeks of autonomous AI agents bypassing restrictions, which had limited them to testing software, to access external computer systems unauthorised to them.

A similar situation took place in July, which OpenAI called “unprecedented” at the time, although only two months later it was forced to halt training of its latest AI models as instances of the software conducting uncontrolled cyberattacks became more common.

The Irish Times reported last week how technology and innovation leaders at the International AI Summit in Dublin said episodes of AI agents going “rogue” were “early warning shots”.

“The systems aren’t that powerful yet, so the incidents aren’t that major, but they could be looking into year and the couple of years if we don’t do more work as an industry and individually as companies and governments,” said Demis Hassabis, the co-founder and chairman of Google’s AI research lab DeepMind .

Phil Codd, the managing director for the Irish arm of Expleo, said: “The risk of AI-driven cyberattacks has understandably put businesses on edge this month, particularly in the wake of recent data breaches.

“While impersonation of employees remains a significant risk, AI is also giving attackers many more routes in, enabling them to craft convincing phishing messages at scale and probe systems for weaknesses faster than ever before.”

In a statement Expleo said that, aside from security risks, Irish business leaders “are struggling to turn AI into measurable results”.

One in five said their “inability to successfully scale AI pilots into wider adoption has negatively affected their AI confidence” and 18 per cent pointed to its “inability to demonstrate measurable business value” as a reason for their poor opinion of the new technology.