DWIC 2022: Is AI the next ‘silent cyber’?

As we stand on the verge of an AI revolution, Simmons & Simmons considered some of the risks for insurance companies using AI in their underwriting practices and the risks in insuring the use of AI.

The challenge is that re/insurers may not be aware of their exposure until it is too late, warned Minesh Tanna, managing associate.

This is due to the nature of artificial intelligence and how it is being used, with vast datasets and highly complex, often opaque, algorithms and advanced analytics sitting behind it.

“Really sophisticated AI systems are so complex that humans cannot understand them,” he said. “AI is operating in a black box.”

The concern is that use of AI may mean insurers and reinsurers are inadvertently falling foul of equality laws, among other things, because of how their systems determine risk selection and pricing.

Tanna anticipates that AI governance will become more prevalent and that regulators will step up their oversight.

“Because of this, organisations are going to have to get to grips with the technology, and the risks of AI.”

Olivia Darlington, head of insurance, MEA, said AI was akin to silent cyber in the sense that many carriers and reinsurers do not yet realise how exposed to AI risks they really are. These exposures have been assumed but not priced for.

Unfortunately, it will take a major loss or series of losses for the industry to wake up to the issue and create bespoke covers, she thought.

Currently, insureds are most likely to try to claim for AI-related losses under their cyber policies, where there is currently the most crossover, although there is also the potential for certain losses to fall within the realm of product liability and professional indemnity insurance.

A hybrid of the three classes is one way the market could seek to affirmatively transfer these emerging risks.