There will still be crashes. There will still be equipment failures- there will still be black ice - there will still be gusts that blow light vehicles-
More than 30000 people are killed on the roads each year. People are comfortable with the current risk level. Self driving cars have to be significantly better, but they will never be perfect and I think most of are okay with that.
Maybe they can get the death toll down an order of magnitude. 3000 deaths per year would be a pretty amazing result.
Maybe the deaths will go up. They’ll face new dangers - malware in their systems- hacking - apparently inadequate safety equipment.
Will someone want to put their baby in something without proper safety? Will people allow safety standards to be lowered beyond what they were 40 years ago? Not likely.
Will lawsuits when someone is injured change manufacturers and stockholders opinions about how unnecessary adequate safety equipment is?
They do it all the time. Every flight is going to have somebody holding a baby in their lap. Buses are driving around and often they don’t even have seat belts. Once the cars are safe enough, people get comfortable quickly.
Think about it this way- we could drop the death rate to 3000 deaths a year next year with a few laws. Drop every speed limit to 25 mph and require are cars to be governed to a top speed of no more than 30 mph. Deaths would plummet but people would be pissed off. For most of us, the benefits of being able to drive faster are worth 25000 lives each year.
Think about it - 3,000 deaths - a fantasy figure pulled out of their asses for pr purposes with nothing to back it up.
Or - really think about it. 6 million car accidents a year. Let’s be kind and say (with no evidence) they’all be 1 million. But the safety equipment sucks so the death rate during an accident goes up by a factor of 100.
Car safety features and their improvements have already significantly lowered death rates. Some car models have had zero deaths. If you want to save lives - increase safety equipment. If you want to kill people- remove safety equipment.
“Getting to zero
The list of models with the lowest death rates illustrates just how much vehicles have improved. Eight years ago, there were no models with driver death rates of zero (see Status Report special issue: driver death rates, April 19, 2007). Now there are nine. These vehicles — which include several luxury models but also some less expensive ones such as the Kia Sorento midsize SUV and the Subaru Legacy sedan — had no driver deaths during the calendar years studied.”
Being merely an order of magnitude safer might not be enough (it certainly isn’t for nuclear power). In fact, there may well be scenarios where improving the safety of robo-cars makes the public feel that they are less safe. Human crashes rarely even make the local news, but a single robo-car crash is headline news around the world. Combined with the innate illusion that being in control makes us safer, I could even see the public resisting self-driving cars to the point that it never fully takes off.
I don’t think the current insurance/traffic law system in the USA works for autonomous vehicles. Not at all.
Under the current system, liability for traffic accidents is distributed across ~220 million drivers and the compensation for those injured/killed is made by the at fault driver and that driver’s insurance premiums. This is acceptable to (almost) everyone since we all share the knowledge that we are fallible humans — any of us could conceivably be at fault in an accident.
In a robo-car world, the “driver” of the robo-car is going to take the position of “who, me?? It was Honda who built the car and Google that programmed the AI. They are responsible for my car plowing into a bus full of tort lawyers’ kids on their way to summer camp, not me!”
And Honda/Google have a lot deeper pockets than any Joe Sixpack driver or his max $1 million payout insurance policy. A couple of dozen accidents could put them out of the robocar business, either directly through gigantic payouts or indirectly through financial risk avoidance.
So for this to work, we are going to need either (1) a clear allowance for possible risks which are not subject to tort law, or (2) a legal ceiling on payouts per victim and per accident, or (3) both. Given the numbers of trial lawyers serving in our legislatures, I’m pessimistic.
The alternative is to learn how to drive. Offer (and require) extensive classes, therapy, lobotomies, whatever it takes. If you can’t figure it out, then you don’t get a license, and either need to move to one of the places in the US where you either don’t need to drive (large cities with good transportation networks) or find other arrangements. Driving is not a right. Just learn to be a functional member of society.
… For certain values of “reasonable”… In my experience, Europe and Asia have usable alternative to the automobile. The United States has a vast array of inadequate band-aid substitutes that only suffice while you’re looking to get another car.
That is seriously wrong. I have no business shaming people over the internet, punitive damages are a work for the judiciary. Lynch law should never be an option.
As this discussion morphs into the pros and cons of self driving cars, I think what most people get wrong is thinking the AI will replace the taxi driver. That’s silly. I think we need to see the AI in a “self driving car” as more akin to a horse pulling a cart — intelligent enough to go around obstacles, follow a set route when the cart driver is distracted, stop when something happens, things like that, but still in need of a driver to tell it where to go, when to go, things like that.
And I agree with both @karl_jones and also @AndreStmaur, that it feels like traffic porn, but also that shaming over the internet just feels wrong. In the end, I think this story should be made public so that people remember what sort of idiots are out there, like the fools who go the wrong way on a motorway (Note that if you are in Germany, A Geisterfahrer auf der Autobahn announcement on the radio means that someone is going the wrong way on the Autobahn)
My wife’s uncle does this. Its because he can’t see very far ahead of him, and it takes him a while to figure out where he is to know if this is where he has to turn. He doesn’t look at signs, relying entirely on the appearance of nearby buildings for navigation.
He is very wealthy and can afford to pay off anybody who gets into a collision with him.
This is why I want to move to the paradigm of treating robo-cars like horse-drawn — the AI is a “dumb beast” which still needs human guidance, and not the superintelligent KITT of our childhood dreams.
Just to Devil’s Advocate for a moment. The film is very short. From that there’s little proof that the driver buggered off after causing those accidents. He pulls his car out off the main road. That’s what he should do in the circumstances. Parking in the highway while he tries to call emergency services and/or getting out to offer assistance would just add to the roadblock.
Kudos though to the driver of the blue truck who ght was going to smash into the guy.
