The idea of artificial intelligence (AI) has fascinated us for decades.
The word ‘robot’ apparently originated in a play called Rossum’s Universal Robots, written in 1920 by Czech playwright Karel Čapek. Robots were creatures that could be mistaken for humans rather than machines.
Notions of a more advanced type of AI came along after the Second World War and led, in 1950, to mathematician Alan Turing conceiving the Turing test: an assessment of a machine’s ability to exhibit intelligent behaviour indistinguishable from that of a human.
AI can have a huge impact on underwriting
As AI came closer to reality, its depiction in pop culture became darker, with films such as Terminator 2 showing it leading to a nuclear holocaust.
Leaps in technology and computing ability over the past 20 years have made genuine AI a reality. But with, it seems, every new tool and system being labelled as AI, what does the term really mean? What can AI do within financial services right now? And where will it lead us?
Independent learning
What AI is not is a series of pre-programmed outcomes based on user choices. Think of a vending machine: it makes an action based on our request, but that is not AI although some fintech providers label it as such.
True AI involves learning, development and growth, which the computer is capable of achieving independently. ChatGPT is one of the best-known examples of proper AI — it learns and improves the more we use it.
In financial services, such AI can play a huge role in enabling us to be much more efficient in our jobs, performing some of the time-intensive tasks that need to be done, from administrative work such as record keeping and note taking to initial client triage.
It seems every new tool and system is being labelled as AI
Our area of expertise lies within insurance, and here AI can have a huge impact on underwriting and pricing. Our decade-long investigation into AI underwriting — to determine how long-established methods that often result in poor outcomes for customers could be improved — has shown that age and other historically driven claims patterns should not be used as key determining factors for price and excess.
Through large-scale analysis of claims data and cancellation patterns, we discovered there were better predictors of risk and anti-selection. A much more accurate way to predict claim propensity and duration is by leveraging global data and employing cutting-edge large language models to analyse, through AI, other lifestyle and behaviour factors. These include everything from information about a person’s job to the mobile-phone handset they use and how many leisure trips they take each year.
By combining multiple layers of real-time data from global data agencies and leveraging detailed regression and correlation models, we find that some clients are deemed lower risk and pay a lower premium than would be the case under traditional underwriting models.
We’ve tested this underwriting against claims data for accident, sickness and unemployment policies and it is currently 80% accurate, which will improve as the AI learns to analyse the data.
AI can play a huge role in enabling us to be much more efficient in our jobs
Eventually, this kind of AI-driven smarter underwriting can be used across financial services to more accurately match risk against price at speed, creating a fairer and more efficient industry as well as giving insurers more certainty that their book of business is well balanced.
There’s no reason why one day this can’t work across all financial services sectors that require risk analysis, from mortgage and income protection to home insurance or even mortgage applications.
Kesh Thukaram is co-founder of Best Insurance
This article featured in the July/August 2024 edition of Mortgage Strategy.
If you would like to subscribe to the monthly print or digital magazine, please click here.
Publisher: Source link
