May Coffee Talk: How will generative AI change scenario analysis?

Originating from Hermann Kahn's work at Rand Corporation in the 1950s to deter thermonuclear war, scenario analysis has matured into an essential business tool, notably within the (re)insurance industry where it is used to strategically handle tail risk. However, scenario science has been a relatively static field of research in recent decades because it is a resource-intensive process that requires foresight, analytical thinking, and imagination. 

A significant challenge insurers face with scenario analysis is the failure of imagination – when bad things happen that were not on our radar, such as the failure of levees during Hurricane Katrina. 

Cameron Rye Head of Modelling Research and Innovation at the WTW Research Network, will discuss how Generative AI, particularly Large Language Models, presents a compelling solution to overcome the limitations of human imagination, while also speeding up the traditional, resource-heavy process of scenario development.

Cameron has over 10 years of experience working in the (re)insurance industry using and evaluating models of natural catastrophes including hurricanes, earthquakes and floods. At WTW Cameron works on a range of research projects, including the evaluation of catastrophe models, development of bespoke views of risk, climate change scenarios, counterfactual analysis and the quantification of non-modelled risks.