Well, the real problem is the drivers who have poor response times for whatever reason. There's no current good way to judge who is really at fault for an accident, so society devolves into witch hunts against drunk drivers, speeding, cell phones, etc. What about simply recording as much data as possible so driver behavior preceding/during an accident can be reconstructed after the fact?
Such a system mandated by insurance/government (same thing) would clearly be an abomination of privacy/ownership issues, but if it was developed for end-users to help prove their innocence, it could be implemented properly (user-based crypto+open formats+local storage).
Why do you argue that members of the public or of an insurance pool have no interest in minimizing the economic cost to them of reckless drivers' risky behavior? Sure, if you drive everywhere at 100 mph and we buy auto insurance from the same company, your direct impact on my premium is minimal, maybe a few cents extra annually, because your increased risk is spread out across a large pool of insureds. But risky drivers exist in quantity too, and collectively their impact on the cost of insurance is large, as is the drag they create on GDP or taxes.
Certainly, I think that vehicle telemetry data should be available to defendants in driving-related criminal cases or litigation. And as a matter of fact, it is, via the legal procedure of discovery. You seem to think that this should only be available to the defendant, but never to prosecutors or plaintiffs. Why not? If you're driving on the public highway, then your driving behavior is a matter of public concern to the extent that it impacts others; and to the extent that other people are exposed to the risks arising from your operation of a vehicle, you cannot reasonably claim an expectation of privacy - unless you're willing to move to a regime of strict liability for accidents, and pay correspondingly higher insurance premiums.
My second paragraph wasn't trying to be philosophical, just what I see as the practical outcomes. If developed for the end user before becoming de facto mandated, there's a much better chance that it would be a system with some notion of privacy between independent actors. A court would have to request the relevant records of the accident rather than getting weeks worth of data ripe with circumstantial evidence. (Whether a defendant can refuse this request is irrelevant, eventually withholding the data while the other party's data shows reasonable actions would be viewed in a negative light)
What would such a system look like if developed by insurance companies? I'm guessing it would contain a cell modem that would continually upload as much data as possible to the insurance company. They'd then datamine that looking for 'dangerous' drivers and charge them more, even if they have less accidents.
Who would you rather have in your insurance pool going forward - someone who speeds everywhere with no accidents, or someone who drives under the speed limit yet has managed to total their car twice in the past year? There's no answer to that - my point is that we just don't really know who the bigger risk is a priori despite much effort, so the game-changing goal should be better analysis after the fact.
Mindlsight, I don't think you have even the haziest idea of how criminal prosecutions or lawsuits work. Courts rarely request anything; prosecutors or plaintiffs make requests through the court, which can be contested by the defense, and the defense can seek records in turn (like the disciplinary record of an arresting officer, for example, or the records of equipment calibration in a DUI case etc.).
There's no answer to that - my point is that we just don't really know who the bigger risk is a priori despite much effort
What do you think insurance actuaries do all day, other than study insurable risk? Read your insurance policy if you have one, they can and will raise your premium if you have a history of accidents. Honestly, you seem kind of paranoid. Insurers want to reduce risk, but still keep their product cheap enough for people to buy it.
I do know how legal discovery generally works, there just wasn't much point getting into specifics when it doesn't change my point. Fundamentally, the court is what has the power to make requests, even though those requests originate from opposing counsel.
> What do you think insurance actuaries do all day, other than study insurable risk?
This is precisely what I meant by 'much effort'.
Let's make this concrete: If you frequented a lot of bars as a designated driver and the insurance companies gained access to your location data, what do you think would happen to your rates? How do you think this change compares to the change that would occur if they could somehow perfectly predict your expected payout?
And actually, yes, I am paranoid and proud of it. The corner cases of systems are the important ones; expected behavior is dead easy.
If you frequented a lot of bars as a designated driver and the insurance companies gained access to your location data, what do you think would happen to your rates?
Nothing, because my location data doesn't tell you whether I'm a driver or a passenger. Even if the insurance companies do have that information, your example depends on their being oblivious to the existence of designated drivers. I might, for example, be operating a taxi or limousine service, where my job is to drive drunken people around while they have a good time.
I'm all about corner cases, but it does not follow that paranoia is the sensible position to take. I can just as easily suggest corner cases where insurance companies decide to skip the analysis, increase premiums on everyone evenly, and blame inflation, because it considerably lowers their overhead and the administrative savings offset a good deal of the increased payouts, so that the premiums don't rise too much. Unsustainable over any period but the very short term, of course, but many would argue that something very similar happened at AIG in the runup to the financial crisis. Not performing proper diligence on the financial instruments whose risk they were insuring against (via credit default swaps) looked like financial genius until the claims suddenly exceeded the value of the premiums. That's why it's cheaper to pay actuaries than to place your faith in people's common sense.
There are different interests to be balanced here, but it seems to me you are complaining about the existence of the scales (ie the court).
The location data of the insured vehicle certainly indicates the context that it operates in. The commercial vehicles you cite have different insurance policies. And letting your friend drive your car home from the bar is presumably an even higher risk behavior! If you honestly think that an insurance company wouldn't raise the rates for something that correlates so strongly with drunk driving, you're really going to need some solid justification.
And as to the scales, one must recognize that they're imperfect. One is in a much worse position with a log of habitual speeding - having to then prove (from a presumption of guilt) that their speed was indeed prudent.
Such a system mandated by insurance/government (same thing) would clearly be an abomination of privacy/ownership issues, but if it was developed for end-users to help prove their innocence, it could be implemented properly (user-based crypto+open formats+local storage).