>As I understand it, the entire concept of insurance is based around gathering relatively large groups of people with a solid understanding of their average ris

>As I understand it, the entire concept of insurance is based around gathering relatively large groups of people with a solid understanding of their average risks for certain types of loss. Given the averages hold accurate over time, those that need a payout can get one, and the insurance company makes money in aggregate. >However, the better the insurance companies can understand our risk profiles as individuals (going well beyond age/sex/smoking/etc/etc/etc) - eventually (in theory) a motivated insurance company is going to be able to determine that I have a 90% chance of developing Serious Syndrome X while you have a 97% chance of developing Minor Issue Y. Even as predictive models improve, insurance will always rely on risk pooling. Even if you believe that insurable events like death are completely deterministic, insurers' information will always be imperfect. So models will always have an epsilon (noise) term, which is diversifiable. >Today we might get put into the risk pool. But as the science advances, the risk pools can be cut into finer and finer (and more accurate) risk pools. Why should a profit-seeking insurance company insure me past some point when the data strongly indicates i'm going to need drastic and expensive intervention. Because they will still profit on average if they charge you more than the expected value of claims. Insurers have historically been hesitant to insure high-risk individuals for any price, in part because they're not confident enough in their ability to accurately quantify the risk associated with them. I'd argue that more granular and accurate rating will actually improve high-risk individuals' access to life insurance for that reason. Edited to add: I suspect that the same would be true of health insurers, though that's currently moot in the US due to the ACA. >It seems like the three ways to avoid this are: 1. "Advances" like the one mentioned here don't turn out to have enough predictive power and insurance continues to operate much as before 2. Some people just can't get for-profit insurance, at least not for health coverage (I'd think accidents will always remain much harder to predict than e.g. heart disease). Pushing them into some kind of not-for-profit 3. Regulation forces blinders on to underwriters essentially saying "actuarial science advances to here and no further" The status quo is mostly #1 for life insurance, and I'd contend that it's likely to stay that way for the foreseeable future. Health insurance is squarely in bucket #3 in the US, but who knows how long that will last or where it will go from here. N.B.: I'm a (non-credentialed) life actuary.