The advance of technology in insurance offers plenty of exciting opportunities. Digitising more of the claims process has the power to greatly improve customer experience. And use of better data up front should allow insurers to shorten question sets – and remove some unnecessary traps that consumers find themselves presented with when they buy an insurance policy today. (Question: Do you have any trees more than 10 metres tall near your property? Answer: How am I supposed to know – my tape measure’s only 2 metres long).
But while digitisation may offer lots of positives, the emergence of Artificial Intelligence in insurance is something which instinctively fills me with fear.
If we are simply talking about removing some of the burden from human underwriters, and allowing complex pricing to become more machine-led, I can see how AI may drive some efficiencies.
But it’s hard to spend more than a few minutes thinking about the use of AI in insurance without reaching a number of more sinister conclusions.
From a consumer perspective, one of the fundamental problems with insurance is that most firms in the market are looking to maximise profit over a relatively short time horizon. That fact instantly pits insurers against the customers that they are looking to serve.
Comparison sites ensure that price competition is intense – leaving insurers forced to make ends meet by cutting cover, putting up ancillary charges, or trying to develop clever pricing techniques that help to attract the very best risks at the maximum premium.
It’s this last thread where AI is most likely to be applied: helping insurers develop an even greater understanding of each risk they take on their book, along with an understanding about that cohort’s sensitivity to pricing changes.
This may deliver better outcomes for a number of consumers. But for every advance in precision pricing, another group of consumers becomes marginalised by the price of their insurance rising. At the extremes, certain groups become uninsurable.
As I’ve said many times before, we’ve moved a million miles from the origins of insurance where customers pooled their risks to protect against unlikely catastrophic events.
Instead, we now have an incredibly sophisticated and complex market in the UK where pricing is designed to represent your own personal risk.
AI threatens to take us ever further down this route – and the end point is not pretty. As I see it, the logical end game is that insurance will become ever more effective at offering competitive deals for those who are least likely to claim, and increasingly intolerant of anyone that sits outside of the conventional risks.
It’s not hard to imagine a world where my insurance price is based on where I shop, what I earn, who I socialise with and what I do in my spare time.
For me, it’s time to put the brakes on such advances. The FCA’s market study, and its consideration of fairness in pricing, is the perfect opportunity to draw some clear boundaries on how insurance is priced. It should start with the principle that insurance should be available for all at an affordable price.
Thankfully, we don’t get taxed according to our life style choices when we go to see the doctor or pay a visit to a hospital. The same universal service obligation needs to apply to insurance too – and I fear that may mean pushing the market back towards its origins. I’m still doubtful that AI has anything very positive to add.
Hear more from James at the ABI Annual Conference 2019 in the breakout session ‘The future of GI pricing: improving customer outcomes in the age of AI’. Find out more and book your place now.