Google is trying hard to fast-track the Get Go of their autonomous vehicles and they are pushing the US government to accept that those “passed federal safety standards”.
This coup is nothing but aiming at turning around liability – should something happen, defendant’s first motion will be to proof it happened despite strictly following that regulation. Any victims might then sue the government, but not the manufacturer of the involved autonomous vehicle.
So unless the regulation is a sound and complete definition of how a vehicle should be operated in all kinds of traffic, hardly a thing anyone came up with for the last 100 years, there is everything to win on Google’s side and not a lot, if anything, on the side of users, today’s human car drivers, mostly.
Of course, that new regulation would apply to human drivers as well. It would be difficult to argue that traffic regulations differ depending on the nature of the vehicle’s operator.
But we need to have that argument. And we need to have the time for it. There is no need to hurry, actually.
Otherwise, we will soon get provisions that limit manual driving on certain roads and at certain times of the day. Eventually, we could be forced to buy autonomous vehicles for our commute and maybe keep the manual one for leisure tours on Sunday mornings.
Google virtually admitted that its car’s software is making guesses about other vehicles’ behavior. That is wrong and at the heart of the problem with autonomous systems.
A human is allowed to use their judgment and make a guess. Eventually, in case of damages, a court will decide whether that judgment was, simply said, good or bad.
An algorithm does not guess. It chooses, based on thoughtful planning and it’s hopefully well-tested design by it’s creators. Now, those are seeking general absolution by the government without having regulators looking funny at those algorithms.
Certainly they will argue that better all vehicles on all roads be autonomous, to prevent an accident as the one they caused and the vehicles would just have negotiated right-of-way.
How much safer would that be! Of course, even better, if all had the same algorithms, for which Google may offer gracious patent licenses.
But, to make that work, we would need to get rid of manual cars. And motorcycles. And bicycles. Deer. Raccoons. Falling trees. Blizzards.
All kinds of road works, too. Those flaggers would need to hold a Bluetooth beacon instead. Or an NFC tag, if they dare. But, really, they can’t be trusted to enter the map changes in real-time, or put up signs with the right QR codes instead of instructions.
Unless, Google is managing all that road works data and sending precise instructions to the workers. The unimaginable increase in security should be worth the comparably small fee associated with that service, I would guess.
And advertisements. Could being driven by Google cars get cheaper the more they slow down or stop in front of billboards? With a customized display of products? And a “OK Google car. Order product to continue trip.” button voice command in the car ?
Individual transport may fall victim to one of the deadliest fallacies of the industrial age: engineers’ (self) over-estimation. They are not inventing with a Plan B in mind. System failure is never an option. Fast Forward. Never in doubt.
Some time ago, in some discussion the question was, “When, do you think, will autonomous vehicles be a matter of course?” And I said, “When Larry Paige let’s a Google car drive his grand-children home from school. During a blizzard.”
So, hold on tightly to that steering wheel. They are coming for it.