But they do use credit score, which negatively affects poor people. They say it is because people with poor credit are more likely to file a claim (note it isn't because of poor driving, just simply ability to cover a smaller incident).
They also use crime rates in an area, how far you drive to work, whether or not you have off street parking - and if that parking is by your house or not. All indicators of a person's finances.
Civil status affects insurance rates - after getting divorced, my rates increased. I've never had a ticket nor an accident. (I think they lower after marriage - somehow, you are assumed to be more responsible even though nothing else changed).
It is amazing how much doesn't depend on your actual driving habits. I understand some of it (ie, crime rate, how much you drive to work) because it does increase risk, but not things like whether or not I'm married.
However, employment status was a very significant factor.
In that sense, it seems as okay as charging poor people a higher interest rate, which seems fine to me.
On the other hand, based on the examples given, it seems this is going to penalise people using language incorrectly - which I would assume can be strongly correlated with poor education. So on that basis, they're going to be saying that poor people are worse drivers and so must pay more? Is that ok?
But the biggest upshot of this whole thing is that suddenly there is a potential for real world financial consequences for what people write online. This is a game changer - and one I would imagine Facebook should be quite worried about.
I hope it triggers the start of a general awakening of the people when it comes to the impact of data overshare.