didds wrote:And there would be an ENORMOUS amount of dead/useless data for those insurance companies' systems to have to weed out..
Crunching data is very cheap these days.
A few years ago I worked on some business intelligence systems and the scary part with those is the stuff they can figure out about people that you just would not predict.
The canonical example given by BI vendors is the link between nappies and beer.
Once they figured out that connection, (young fathers being sent out to buy nappies and deciding to pick up some beer at the same time), many supermarkets started moving the beer and nappies so they are together, either on the next aisle or right on the next shelf, and sales of beer shot up.
Less pleasant example is that they can predict a large percentage of fraud just by looking at peoples other activity. The problem is that these things are never 100% accurate, but are sometimes treated like they are factual rather then an imperfect prediction.