AI Driven Personalized Life Insurance Premiums; Aye Or Nay ?

The mortality risk is not the same across different sections of the population.So one of the ways in which life insurance companies have traditionally competed is by restricting their offerings to people who have lower risks of dying early and thus keeping premiums lower and/or bonuses higher.The classic example is the Postal Life Insurance plans of the past that were offered exclusively to government servants vs similar plans of the LIC that were open to all. In more  recent times,insurance companies target the more educated, affluent, urban ,financially successful professional/managerial class with better living standards and access to world-class healthcare.
Insurtech powered by AI can increase the ability of life insurance companies to make fine distinctions in mortality risks, resulting in more risk classes .So one of the merging trends in life insurance is the increasing number of risk classes as insurers seek to better match risk with premiums.Thus in theory, rather than dozens of risk classes, thousands of risk classes or even lakhs of risk classes are possible. In fact if the AI is astute at risk classification, its possible that each policyholder would be in his/her own risk class. In other words, we would have personalized premiums,designed and calculated on a custom basis for each insured person.
This is a revolutionary thought indeed, as with personalized premiums there will no longer be any risk pooling or risk sharing. And all of the actuary’s toolkit relating to group-average statistics that have served so well since 1693  when Edmond Halley published his An Estimate of the Degrees of Mortality of Mankind, Drawn from the Curious Tables of the Births and Funerals at the City of Breslaw, with an Attempt to Ascertain the Price of Annuities upon Lives would be replaced by individual-based predictive analytics.
Now would I buy such personalized insurance products ?
Rather than a straight Aye or Nay ,my answer is Depends.AI or not,the first principles of life insurance have not changed.Life insurance products exist to help people deal with the issues that rise from the 2 distinct possibilities of their life.The first ,an early death before they have discharged all their duties to their loved ones and secured their material interests.The second,a long life that stretches perhaps decades beyond retirement, to live which with financial independence and dignity, a person needs assets they can draw upon.Additionally,many desire to leave behind a legacy,secured by a life insurance policy.So keep these factors in mind even as you evaluate newer kinds of Life insurance products to see how well they serve your needs.
Now,the AI I’ve been playing round with predicts that I’ll live even longer than my astrologers predicted.A whole decade longer.And both these give me more years than the national average reflected in the actuary tables used by insurers.So,perhaps personalized premiums would be lower than average for me,if the proprietary AI used by various companies predicted something along the lines of the AI I use.If companies had enough faith in their AI to not pad up error margins .And companies didn’t simply use AI to increase their own profits,leaving premiums untouched.
But should personalized premiums be lower,would I buy ?Most definitely,in case of term plans as long as I see the company as having a good record of promptly honouring genuine claims.I case of ULIPs I would need to clarity on how the lower mortality risk would be reflected in the fees/deductions etc.Till date,only one insurer has promptly given me the details of how they would price risk in my ULIP purchase and stuck to their promise.The rest have engaged in glib talk and resorted to stonewalling in the face of persistent questions.So if I perceive ULIPS drawing on AI as making the power relations between the insured and the insurer more skewed,I’d be more inclined to give them a miss.
Till date,most of my insurance purchases have been some variation of the endowment plan,mostly of the whole of life variety,those with relatively predictable,though not necessarily guaranteed returns. What will happen to this class of insurance products in future,I’m not quite sure.Over the last decade,I’ve see landmark plans withdrawn and never replaced with something similar.And sadly the newer alternatives introduced are simply not as great investment-wise as the withdrawn ones.I’m not sure if this withdrawal is a result of changes in the financial outlook of the world or a response to consumer preferences.I suppose,as and when I’m offered a personalized endowment plan,I’ll still be reaching out for a spreadsheet to cue in the inflows and outflows and try to work out rates of returns and comparing with the alternatives then available to see if it looks like a good deal for the years ahead.
In India, we are used to having premiums fixed throughout the premium paying period of our plans.Some ULIPs and the rare endowment plan allow for the insured to voluntarily pay more that goes towards the savings and investment component of their plans.Elsewhere in the world there is however something called the variable or adjustable premium.Perhaps in future, life insurance companies using AI drawing on some periodic or continuous monitoring of our bodies and habits might offer us plans with variable premiums.Would I consider buying them ?No.The reason is simple. I buy insurance products  to bring about greater certainty and predictability in my life so a variable premium that I could not predict in advance and include in my calculations before purchasing the product would simply not do.
Life insurance companies often sell annuities too.In case of annuities,AI could reverse the dynamic.Those who could expect to pay lower life insurance premiums because AI predicts a long life for them,might have to deal with smaller payouts from their annuities as their corpus would have to last longer without the danger of running out.Assuming of course that insurance companies can’t find a way to get and pass on better returns from money that could be deployed for longer periods.In case of annuity plans that call for accumulating a corpus before payouts,if AI predicts a longer life,there is a chance that personalized premiums could be smaller. Having bought an plan to build an annuity once before,would I buy another?Especially if AI made premiums lower ?The returns offered were interesting?Perhaps.As of today annuity payments are taxed,unlike the survival benefits of life insurance policies.When I have no way of knowing what will be the tax rates a few decades down the line,I wouldn’t like to load up on a class of products whose after tax returns might make my older self not so happy with my younger self.
So overall,for the more fortunate of us,AI in insurance is probably going to deliver atleast a few good things to look forward to.But I wonder what happens to the less fortunate of us?Those that AI might judge as not the best mortality risk ?What when they are not necessarily well-to-do people who can overlook the higher premiums,same as smokers and adventure sport enthusiasts do with the loading they have to currently bear?Just as we talk Universal Basic Income to address the problem of jobs and incomes displaced by automation will we have some sort of AI-Backed Universal Basic Coverage for those who can’t afford AI driven personalized life insurance ?For there is always the danger that if insurtech is seen as acerbating the differences between the Haves and Have Nots,society might end up saying Nay to AI driven personalized life insurance.