Autonomous Cars

[Greg Locock]

All the big OEMs have said they will be liable for damage caused by their true AVs. And that is why the big OEMs are not releasing half arse L2 or L3 cars that will hand control back to the driver with little notice. They have them, I've been driven by one, but they are not used on public roads other than for research and testing. I do not understand the Uber case result. I suppose if Uber accepted responsibility and paid off the family to their satisfaction then the letter of the law has been met, but somebody, somewhere, killed someone and should not just get off with a financial penalty. That is a more general problem with corporations and the law. Kill a few hundred people and pay some compensation.

[Phil]

Just wait till the first people are killed in accidents and the discussions will inevitably reach another level about responsibility, liability, collateral damage etc. it’s one thing to hold another human responsible for a mistake, but entirely something else to hold a coorporation, software, machine responsible for the same. And it will come. Just wait until more of these things start driving around without supervision and entirely in the hands of customers.

The expectation is for it to be safer. That is a prudent assumption for any inherently conservative large enterprise. Consider that airline crashes rarely end corporations. They will be aiming for airline levels of incident rates. Tesla will offer their own insurance for this reason. Reduced average incidents rate will offset mortality rate, which should be reduced as well. Cheap insurance that can still pay off the large incidents. Tesla have weathered all the fire and explosion press. Sensationalism hasn't detered them. Same will apply to autonomy.

Usually when people die in airplane accidents, they took the risk of flying. What I am talking about is not people driving an AV being at risk - I'm talking about incidents where an innocent by-stander, a pedestrian, a fellow motorist is killed or tragically injured.

And most airplane crashes are indeed due to human error or extra ordinary circumstances. Planes are also not autonomous. There's a pilot in there for a job and for every eventuality. Do not compare this to an autonomous vehicle driving on roads with thousands of unpredictable factors. Flying a plane in an (relatively) empty sky is a piece of cake in comparison compared to the amount of processing power that is required to deal with driving on a road with an unpredictability waiting to happen every second.

[Big Tea]

Just wait till the first people are killed in accidents and the discussions will inevitably reach another level about responsibility, liability, collateral damage etc. it’s one thing to hold another human responsible for a mistake, but entirely something else to hold a coorporation, software, machine responsible for the same. And it will come. Just wait until more of these things start driving around without supervision and entirely in the hands of customers.

The expectation is for it to be safer. That is a prudent assumption for any inherently conservative large enterprise. Consider that airline crashes rarely end corporations. They will be aiming for airline levels of incident rates. Tesla will offer their own insurance for this reason. Reduced average incidents rate will offset mortality rate, which should be reduced as well. Cheap insurance that can still pay off the large incidents. Tesla have weathered all the fire and explosion press. Sensationalism hasn't detered them. Same will apply to autonomy.

Usually when people die in airplane accidents, they took the risk of flying. What I am talking about is not people driving an AV being at risk - I'm talking about incidents where an innocent by-stander, a pedestrian, a fellow motorist is killed or tragically injured.

And most airplane crashes are indeed due to human error or extra ordinary circumstances. Planes are also not autonomous. There's a pilot in there for a job and for every eventuality. Do not compare this to an autonomous vehicle driving on roads with thousands of unpredictable factors. Flying a plane in an (relatively) empty sky is a piece of cake in comparison compared to the amount of processing power that is required to deal with driving on a road with an unpredictability waiting to happen every second.

The plane, or wreckage still falls where it will. There is a town in Scotland called lockerbie, and I an sure many more.


https://en.wikipedia.org/wiki/Pan_Am_Flight_103

There is also a fear of deliberate interference such as the world trade buildings.

[Phil]

Do you really want to compare people killed by airplane debris with the sheer number of people affected by accidents on public roads? Really?

[subcritical71]

When assigning liability to an AV, with all the unknowns on the roadway, wouldn’t you also need to look at what caused the accident? In other words, if a pedestrian was crossing outside of a cross walk and initiated the accident shouldn’t they be the ones held responsible? It’s not the AV’s (or AV designers) fault that some outside influence caused an accident anymore than if the same would happen to a human driver.
In my mind, AV's will never have the human capability of improvisation, but I still believe that AV’s in general will be safer and should not be subject to over reaction in an accident just because it was an AV. Even human improvisation can make an accident worse by taking the same ‘superior’ improvised reaction that turned out to be the wrong course of action.

[roon]

Just wait till the first people are killed in accidents and the discussions will inevitably reach another level about responsibility, liability, collateral damage etc. it’s one thing to hold another human responsible for a mistake, but entirely something else to hold a coorporation, software, machine responsible for the same. And it will come. Just wait until more of these things start driving around without supervision and entirely in the hands of customers.

The expectation is for it to be safer. That is a prudent assumption for any inherently conservative large enterprise. Consider that airline crashes rarely end corporations. They will be aiming for airline levels of incident rates. Tesla will offer their own insurance for this reason. Reduced average incidents rate will offset mortality rate, which should be reduced as well. Cheap insurance that can still pay off the large incidents. Tesla have weathered all the fire and explosion press. Sensationalism hasn't detered them. Same will apply to autonomy.

Usually when people die in airplane accidents, they took the risk of flying. What I am talking about is not people driving an AV being at risk - I'm talking about incidents where an innocent by-stander, a pedestrian, a fellow motorist is killed or tragically injured.

And most airplane crashes are indeed due to human error or extra ordinary circumstances. Planes are also not autonomous. There's a pilot in there for a job and for every eventuality. Do not compare this to an autonomous vehicle driving on roads with thousands of unpredictable factors. Flying a plane in an (relatively) empty sky is a piece of cake in comparison compared to the amount of processing power that is required to deal with driving on a road with an unpredictability waiting to happen every second.

Not trying to compare the operations or parameters of the two, merely saying airline incident rates or similar is the goal for AVs i.e. low rates generally i.e. safety. Substitute any low frequency occurence for airlines in my previous post.

[Big Tea]

Do you really want to compare people killed by airplane debris with the sheer number of people affected by accidents on public roads? Really?

The insurance companies would, they would have to.
An AV driving into say a marching band or parade, or a fuel tanker all have to be considered no matter how low the probability. It is very very low risk a piece is going to fall of a plane and land in a crowd, but the insurance companies have to consider it.

On a prime flightpath there may be a plane every few min, but on a busy street hundreds of cars per hour will pass so the risk is many times as much even if the damage will be many times smaller.

As I said above though, it will be scare mongering until the novelty wares off and it becomes an every day thing.
It is not GOING to happen but you can bet the press will say it CAN happen.

[Phil]

The insurance companies would, they would have to.

Hence statistics and probability. Planes fly in the sky, in designated air routes. At the altitude they fly for the majority of the flight, the risk is extremely minimized to encounter severe unpredictable dangers. Weather is a factor, but for the most part avoidable. The biggest problems with air travel are related to safety and security which can be countered by maintenance, high safety and security standards which makes air travel extremely safe by all measures. When a plane crashes (which is rare in itself), it rarely affects people on the ground.

The sheer number of cars on the road (increasing), the density (also increasing) and the high number of unpredictability one might encounter makes driving increasingly dangerous. Compared to that flying is a drop in a bucket, so yes, insurrance companies are well informed.

AVs will not counter that. AVs will not reduce the increase of vehicles on the road. It will not reduce traffic density. Accidents will happen and when they will, they will also impact innocent bystanders. When they do, it will come down to the question if this is simply collateral damage in the progress of technology or if we are creating new problems in trying to solve others.

The only thing AVs have going in their favor is that it's a glorified self-driving taxi. Not much more. I really find it fascinating that people think this is something that will benefit humanity in some way. The only line of logic I can somewhat follow is that by enforcing AVs to take over manual-cars is that the human element is removed - thus the 0.00001% of people who are a risk on roads are replaced by a piece of software. Accidents by humans in cars are IMO not a result of people being "bad" at driving. Some may be influenced by alcohol and drugs and more increasingly distractions at the wheel (conversation, phone, texting, navigation, smoking or other distractions), but I also see the increase in accidents as a net result of increasing traffic density on roads.

If safety was the only real concern for the push for autonomous vehicles - why not simply push public transportation in automated trains and subway for all I care? Much smarter in light of environmental concerns too btw. Oh but wait, no car manufacturer has a vested interest in public transport. Care to guess why?

[strad]

if a pedestrian was crossing outside of a cross walk and initiated the accident shouldn’t they be the ones held responsible?

Pedestrians have the right of way in or out of a crosswalk.

[Phil]

IMO it's a good point where liability is concerned. Should an AV be more liable if an accident occurs if a pedestrian (e.g. a kid) runs out onto the road? Maybe or maybe not. It's all in the numbers. I just think a piece of software should be subject to a higher level of scrutiny than a human, because in the end, a human can be held accountable while a piece of software can not.

[hollus]

...if this is simply collateral damage in the progress of technology or if we are creating new problems in trying to solve others...

Creating new problems while solving other is the very nature of the progress of technology. Hopefully the new ones are lesser problems.

...thus the 0.00001% of people who are a risk on roads...

That is one in 10 million. But you probably mean "a very low number". Well... do you know anyone that has been involved in traffic accidents? Because I know several people and I don't know that many people in total.

...why not simply push pubic transportation in automated trains and subway for all I care?

I live in Copenhagen and I have been riding a driver-less, fully automated subway since I got here.

[Big Tea]

The insurance companies would, they would have to.

Hence statistics and probability. Planes fly in the sky, in designated air routes. At the altitude they fly for the majority of the flight, the risk is extremely minimized to encounter severe unpredictable dangers. Weather is a factor, but for the most part avoidable. The biggest problems with air travel are related to safety and security which can be countered by maintenance, high safety and security standards which makes air travel extremely safe by all measures. When a plane crashes (which is rare in itself), it rarely affects people on the ground.

The sheer number of cars on the road (increasing), the density (also increasing) and the high number of unpredictability one might encounter makes driving increasingly dangerous. Compared to that flying is a drop in a bucket, so yes, insurrance companies are well informed.

AVs will not counter that. AVs will not reduce the increase of vehicles on the road. It will not reduce traffic density. Accidents will happen and when they will, they will also impact innocent bystanders. When they do, it will come down to the question if this is simply collateral damage in the progress of technology or if we are creating new problems in trying to solve others.

The only thing AVs have going in their favor is that it's a glorified self-driving taxi. Not much more. I really find it fascinating that people think this is something that will benefit humanity in some way. The only line of logic I can somewhat follow is that by enforcing AVs to take over manual-cars is that the human element is removed - thus the 0.00001% of people who are a risk on roads are replaced by a piece of software. Accidents by humans in cars are IMO not a result of people being "bad" at driving. Some may be influenced by alcohol and drugs and more increasingly distractions at the wheel (conversation, phone, texting, navigation, smoking or other distractions), but I also see the increase in accidents as a net result of increasing traffic density on roads.

If safety was the only real concern for the push for autonomous vehicles - why not simply push public transportation in automated trains and subway for all I care? Much smarter in light of environmental concerns too btw. Oh but wait, no car manufacturer has a vested interest in public transport. Care to guess why?

The way I see it, and it is only my view, is the biggest advantages are when there is no one even in the car.
A strange way of looking at it? Maybe, but the biggest problems come what you have finished your trip, and with electric, getting it charged.

If you can get out of the car, and it goes out of (say) town center, goes to a parking spot and charges its self up until you call it again, think of the reduced congestion and freeing up room that is only used to put a car when it is not even needed?

If it ever becomes practical to have a 'car share' system, even better, you can then have half the number of cars.

[Andres125sx]

I think anyone purchasing an AV will have to assume the responsibility, at least while the car has a wheel and you can drive it. If you want to let the car drive itself it´s up to you, but the responsibility is still yours

Once AVs are so developed they don´t need a wheel, things will change, and probably even manufacturers will start assuming responsibility as a marketing strategy...

Our cars are so safe we assume responsibility for any accident Because they know the rate will be so low the sales increase will be worth. Same as some companies assuming return costs of their products if you don´t like it after purchasing: "give it a try and we´ll return you the money if you don´t like it, we´re this sure you´ll love it"

I think with AVs this will be the normal process

[Phil]

...thus the 0.00001% of people who are a risk on roads...

That is one in 10 million. But you probably mean "a very low number". Well... do you know anyone that has been involved in traffic accidents? Because I know several people and I don't know that many people in total.

I guess the number i was trying to come up with is so small, it’s not quantifiable. Think of it like this: even a dangerous irresponsible driver will not always be a risk to others during the tens-of-thousands of kilometers he might cover in his lifetime. On the other hand, a perfectly responsible driver might find himself distracted for 2 minutes on his 2 hour commute as he is texting someone while driving. 9999 out of a 10000 times nothing might happen. Or he might strike gold and something happens that could have been avoided if only he’d not be distracted at that very instant. The point being, even dangerous drivers or drivers who are easily distracted may drive thousands of kilometers without having an accident or being a substantial risk to others.

And yet, by the sheer number of cars on the road, the immense density of traffic, accidents do happen quite regularly simply as a result of the pure quantity of them. Right now AVs are deemed as “safe” because there is only one of them among... a million of non-AVs? If the software isnt up to it and every single eventuality, i’m quite certain they have a long way to go until they catch up to even distracted humans, statistically.

[Big Tea]

Another point that may be worth considering is that some people seem th have had several car accidents, while most with plane accidents have only one.

(just kidding really, but it does affect the statistic)