And thats where it gets messy and infringes on personal freedoms: if you're required to purchase a risk-adjusted product but you don't know what risk factors you're being judged on, how do you know you're not being discriminated against unfairly? How do you know there's no mistakes? How do you know the person you sold your car to isn't accidentally affecting your rates because the insurance company's data broker never got the transfer-of-sale? How do you know the insurance companies aren't colluding and deciding you're on the wrong side of a Red Line they drew on some map?
With credit scores, there's been a long fight to get a half-assed kinda-free basically-complete credit report so you could at least see what you're being judged on. Where's the "credit report" my insurance company is making choices on so I can at least try to correct inaccuracies?
I get insurance companies make decisions based on information, and better information is just market efficiencies at work. It gets real messy real quick when the information affects something you're often required to purchase and you have no recourse.
Go ahead, get the market efficiencies from better information, but if you're being forced to participate, the government forcing your participation should also force the other side to present the information that judges you.
Ignoring the fact that reckless driving is an actual crime and thing they did, not "something they didn't do", it is also probabilistically the thing that leads to collision damages. Insurance companies are probability companies first and foremost. It absolutely makes sense for them to charge more to cover behavior that is more likely to lead to payout.
I'm not a fan of the surveillance without consent, but if this were opt-in I would be ok with reckless drivers being charged more for the wake of accidents they may leave in their path.
I agree that it smells terrible.
But insurance is full of this kind of thing -- it's about predicting possible outcomes, identifying the random component that goes into it, and charging based on likelihood.
Consider the example of life insurance companies charging smokers more, even if they don't have emphysema or lung cancer yet. Maybe they should quit smoking but we understand the connection between risk factors and outcomes well enough that it's not unreasonable to charge more.
(Not exactly comparable, of course: it's hard to argue "well, I've smoked for 30 years and so far so good!" in the same way as "I brake hard a lot but I've been doing it for 30 years without an accident so clearly it's not a problem".)
A big problem here is lack of consent. It's one thing to add an insurance surveilance device to your vehicle, it's another for it to be built in and not specifically agreed to.
Another big problem is lack of transparency: we don't always know what they're tracking.
Another big problem is (most likely) lack of understanding of the scene. You can't tell from one car's sensors if the driving was appropriate, IMHO. Maybe someone with a lot of hard braking should drive differently, but if they haven't had any collissions and have a long history, it smells of 'pre-crime' to charge them more for something they didn't do.