ARTIFICIAL intelligence is rapidly shifting from a technical novelty into an immediate operational hazard, rapidly upskilling threat actors and forcing organisations to overhaul their approach to risk management.
National Risk Solution director Chris Graham made these comments while speaking at the Pacific Fiber Conference 2026 in Nadi, saying brokers and underwriters across global hubs from London to Singapore had reported that cyber exposures now reached into hundreds of millions of dollars.
The core issue, experts warn, is that tools once locked behind months of specialist dark web collaboration are now broadly accessible.
Mr Graham said AI was effectively arming unsophisticated criminals overnight, accelerating deceptive tactics like targeted phishing and automated fraud.
At the same time, he said companies were eagerly embedding AI into internal workflows to capture efficiency gains.
Yet, he added that rapid adoption had created dangerous blind spots around unapproved employee software, data transfers, and rogue automated agents.
“Risk management cannot remain a once-a-year exercise, cyber risk is now a business loss issue, not only a technical incident,” Mr Graham said.
“It can become downtime, lost money, legal cost, reputational damage, and recovery expense,” he said.
While human error and everyday scams remained frequent drivers of claims, Mr Graham said high-severity events like ransomware continued to dominate organisational losses.
Modern insurance covers must now serve as active resilience tools funding threat monitoring, forensics, and crisis response rather than acting as passive payout policies.
To withstand evolving threats, Mr Graham proposed that leaders must regularly test internal controls, establish clear accountability for AI failures, and present verifiable risk management practices directly to underwriters before severe market losses occur.
