The AI risk the UK is least prepared for

Insurers have warned that AI-enabled cyber attacks are an emerging risk that the industry is least prepared for in the coming years.

According to a survey by KPMG, 30% of UK insurance leaders believe that AI-enabled cyber attacks as the emerging risk the industry is least prepared for. The risk was ranked above that of climate-related catastrophes or technical failures and outages.

Frontier AI models are increasingly being leveraged to expose security vulnerabilities in software and protocols used across the internet and enterprise infrastructure. Just this past week, a critical security vulnerability was detected in PHP, which is endemic across web infrastructure, that allowed attackers to deploy and run code remotely on target web servers.

Additionally, a critical vulnerability was detected in GitLab’s Community and Enterprise editions this past week, prompting a global response as businesses scrambled to secure their infrastructure.

A widely-used Indian certification authority was also compromised this week after it’s private keys were found to be far more unsecure than designed, with browsers and infrastructure providers scrambling to untrust the poisoned certificates.

In most, if not all, of these cases, it can be presumed that AI models were used to hunt for vulnerabilities, as they are proving extremely effective at code-related tasks, specifically around cybersecurity.

With attacks on widely adopted infrastructure mounting, UK insurers are beginning to apprehend the potential consequences of this technology on their business.

According to KPMG’s survey, six in ten insurance leaders say cyber resilience should be strengthened over the next 12 months, and more than a quarter believe that cybersecurity should be the UK’s top priority for strengthening national resilience over the next year, more than conventional defence.

“The rapid development of AI frontier models is challenging insurers’ operational resilience against a range of rapidly evolving cyber attacks,” said KPMG Financial Services Head of Insurance and Partner Huw Evans.

“As concerns around the risks of AI frontier models intensify, insurers are increasingly acting at speed to enhance protection, strengthen cyber protection and use advanced AI to protect themselves against AI-enabled threats.”

“Cyber risk also increasingly sits within a much wider geopolitical picture. As an industry, we need to improve our ability to translate geopolitical risk into better underwriting, risk management and overall resilience,” he said.

Now read: Pay or be tracked – Big changes coming to Yahoo in the UK