Brought to you by:

‘Urgent action needed’: frontier AI threat ranks high on APRA agenda

The prudential watchdog fears “more severe disruptions” from frontier artificial intelligence and has pressed insurers and other regulated entities to quickly strengthen risk management.

The warning comes in the Australian Prudential Regulation Authority’s 2026-27 corporate plan.

On risks specific to general insurance, APRA says affordability and availability remain under pressure amid high claims costs, natural hazards and climate-related risks, contributing to a widening protection gap.

“It’s a very big structural issue in the general insurance area,” chair John Lonsdale said this morning.

“From APRA’s end, it is not something that we look to solve, and nor can we solve by ourselves. It’s really a structural issue for regulators, for government, for the private sector and for individuals actually to lean in to.”

He says the authority’s recent work on insurance challenges includes its inaugural climate vulnerability assessment, stress testing householders’ protection gap. Findings released in March showed more homeowners may be forced to go without property insurance by 2050.

“One of the important results out of that [stress test] was, when you look at underinsurance at the moment … [it] is sort of talking about one in seven households under a plausible stress scenario,” Mr Lonsdale said.

“That figure could – it’s not would – could worsen to one in four, and that’s a very significant development.

“What we tried to do in that report is set out what we’re doing, which is really putting information on the table, analysis on the table, providing our expertise through things like the [Hazards Insurance Partnership], and also feeding the results into good risk management practice with our regulated entities.”

The corporate plan setting out APRA’s strategic priorities for the next four years notes “new and more complex challenges lie ahead”.

Rapid technological developments such as frontier AI models could create more severe disruptions.

“During 2026-27, we will focus on ensuring entities strengthen their resilience to AI-enabled cyber threats,” the authority said.

“We expect entities to maintain effective cyber controls as threats evolve, improve board oversight of technology and cyber risk, and participate in industry information-sharing where this can support broader system resilience.

“Rapid advances in frontier AI capabilities require an urgent strengthening in technology and cyber resilience.”


From the latest Insurance News magazine: Why the "patchwork" approach to home insurance affordability just won't cut it