3 comments

[ 6.1 ms ] story [ 102 ms ] thread
I'll take "whoever is legally responsible" as my answer.

Did the car crash because of faulty programming? Then I imagine the car manufacturer is liable.

Did the car crash because it was hit by a drunken idiot? Then the drunken idiot would be my guess.

Some combination of the above? Then some combination of responsibility is assigned.

And so on.

It will vary on a case by case basis while the law discovers the common permutations of facts in automated car crash cases. Or until legislatures decide as a matter of policy to introduce a common scheme.

I'm assuming we're not talking about "no fault" jurisdictions. Because in no fault schemes you just skip the civil question of fault altogether and go directly to the question of fixing things up. (There might still be criminal charges). No fault isn't without problems but it certainly saves a lot of court time and money.

(IANAL, TINLA)

#2 and similar situations will be _well_ mitigated by the plethora of data collected before and during the accident. It's no longer a question of who-said-what, where the impact was on the car, and estimates based on skid marks.

Between recorded lidar imaging, video imaging, and all the telemetrics needed to automatically drive the car:

Driverless car impacting other driverless car; probably a software error.

Driverless car impacting / being impacted by a piloted car, and it will be fairly easy to see which driver made the error which led to the crash.

Car / AI manufacturers will have more of a liability for making certain their vehicle is safe, no doubt. This is a win/win situation for the automobile owner. Safety is now an economic force not dictated by regulation.