Moody’s RMS’ Steel: Generative AI is a real game-changer for the industry
Generative AI offers a “huge amount” of capabilities that can help enhance modelling and event response, as well as deliver process improvement for the wider industry, according to Moody’s RMS general manager Mike Steel.
In an interview with The Insurer TV, Steel noted that AI has played a role in model development for the past decade, but said its usage accelerated – particularly for event response – during the pandemic.
“We couldn't put people on the ground to go and look at damage within buildings. So we utilised AI technology to assess satellite images and generate our own loss estimates at a faster pace.
Steel said a “huge amount” of capabilities can be delivered through generative AI, with the modelling firm recently agreeing a partnership with Microsoft that facilitates the use of generative AI in intelligent risk platforms to deliver targeted data for clients
“If there is a hurricane off the coast of Florida, you can interrogate the AI by asking questions like what's the likely loss from this type of hurricane? Were there similar hurricanes that have happened in the past with a similar type of footprint? What contracts would be affected by this hurricane,” explained Steel.
While the process would in the past have taken days or weeks, Steel said generative AI can deliver the answers in seconds.
Secondary perils “no longer secondary”
Increased losses related to what have traditionally been dubbed secondary perils – such as wildfires and severe convective storms – have Steel wishing the industry would banish the phrase from its lexicon.
“I’d really like to expunge the term secondary perils from the dictionary if I could,” he said. “Floods, wildfires and severe convective storms are increasing in frequency. We're seeing those types of perils with a very strong climate signature increasingly hurting the earnings of companies and causing some material losses,” he said.
“In fact, if you look over the last, say, five or so years in the US, and count up the level of severe convective storm losses, they're probably equal to the amount of windstorm losses that we've seen, they're that impactful,” said Steel.
He said Moody’s RMS is now investing heavily in high-definition models that are better able to capture the nature of these types of localised threats, to study the risks on a more granular basis. Moody’s has also added the option of climate risk scores, which consider additional factors like heat stress and drought.
The added perils have changed the dynamic for insurers, according to Steel. It’s no longer enough, or even possible, to just get out of the way of risk.
“They should understand the risk, figure out how to underwrite it, and then provide those products to society,” said Steel.
Steel said Moody’s has seen an uptick in interest from insurers, regulators and rating agencies to understand the impact of climate change, as well as from banks and asset managers.
“The fact that the climate risk scores that we have and the climate condition models are based on exactly the same science, with the same people within RMS driving that science
means that we've got a common currency of risk between those banks, asset managers, and corporates, and the insurers, so when an insurer engages with a corporate about climate change, they're talking the same language,” said Steel.