LA TIMES: Walt Disney World plans to deploy driverless shuttles

[MisterPenguin]

Yes.

That doesn't mean Disney would be willing to take 100% liability and dropping these on roads with 40-50 MPH speed limits...

Why do people keep getting hung up on liability as if their crop of human bus drivers do not expose them to the same liability? Disney is liable whether their human drivers get in an accident or their autonomous vehicles get in an accident. That's why there's automotive insurance, to cover that liability. Someone who is harmed by an autonomous vehicle isn't getting a larger insurance claim payment than if it were a human bus driver who caused the accident.

If autonomous vehicles are significantly better than human drivers, even by 10%, then the liability exposure goes down, not up. Less accidents, less damage, less death. Less insurance pay-outs, lower premiums.

 

[GoofGoof]

Why do people keep getting hung up on liability as if their crop of human bus drivers do not expose them to the same liability? Disney is liable whether their human drivers get in an accident or their autonomous vehicles get in an accident. That's why there's automotive insurance, to cover that liability. Someone who is harmed by an autonomous vehicle isn't getting a larger insurance claim payment than if it were a human bus driver who caused the accident.

If autonomous vehicles are significantly better than human drivers, even by 10%, then the liability exposure goes down, not up. Less accidents, less damage, less death. Less insurance pay-outs, lower premiums.

The vast majority of autonomous vehicles in testing today still have drivers there in case the system fails, most places require it. In samples I have seen with driverless shuttles in actual use (not testing) the vehicles go pretty slow (not highway speed) or run on dedicated roadways. Assuming everything works 100% the way it should you may be right that the driverless vehicle may be 10% or more safer than a human driver, but if everything doesn't work exactly right and there's nobody on board to take emergency action that's when there's a problem and where the extra liability comes in. The other thing autonomous vehicles cannot do as well is use judgement or interpret things like hand signals from emergency workers. They also cannot make "moral" decisions. If a car is cutoff and an accident is unavoidable we rely on a human driver to make a decision to either hit the car in front of them or swerve into the next lane or shoulder and potentially hit the car next to him. They make a split second judgement call as to which action is less negative. Those judgement calls are not always going to be correct, but you can't program a computer to make that kind of decision since everything is rules based and both outcomes would violate a rule. The computer can't pick the lesser of 2 evils. I think that's the main reason that in practice these systems tend to go slower and in more controlled environments.

 

[Nubs70]

Why do people keep getting hung up on liability as if their crop of human bus drivers do not expose them to the same liability? Disney is liable whether their human drivers get in an accident or their autonomous vehicles get in an accident. That's why there's automotive insurance, to cover that liability. Someone who is harmed by an autonomous vehicle isn't getting a larger insurance claim payment than if it were a human bus driver who caused the accident.

If autonomous vehicles are significantly better than human drivers, even by 10%, then the liability exposure goes down, not up. Less accidents, less damage, less death. Less insurance pay-outs, lower premiums.

Why,?

Because the legal system has not yet defined liability in the case of autonomous vehicles.

 

[MisterPenguin]

The vast majority of autonomous vehicles in testing today still have drivers there in case the system fails, most places require it. In samples I have seen with driverless shuttles in actual use (not testing) the vehicles go pretty slow (not highway speed) or run on dedicated roadways. Assuming everything works 100% the way it should you may be right that the driverless vehicle may be 10% or more safer than a human driver, but if everything doesn't work exactly right and there's nobody on board to take emergency action that's when there's a problem and where the extra liability comes in. The other thing autonomous vehicles cannot do as well is use judgement or interpret things like hand signals from emergency workers. They also cannot make "moral" decisions. If a car is cutoff and an accident is unavoidable we rely on a human driver to make a decision to either hit the car in front of them or swerve into the next lane or shoulder and potentially hit the car next to him. They make a split second judgement call as to which action is less negative. Those judgement calls are not always going to be correct, but you can't program a computer to make that kind of decision since everything is rules based and both outcomes would violate a rule. The computer can't pick the lesser of 2 evils. I think that's the main reason that in practice these systems tend to go slower and in more controlled environments.

First of all, humans have to make the same moral judgements in a split second and they're generally bad at it. Everyone asks what an automaton would do if having to chose to hit and kill a pedestrian to save the driver, or, swerve to save the pedestrian which would kill the driver. The first question is: what do human beings do in that situation? What is the correct thing for a human driver to do? Having automatons doesn't create that dilemma, it forces us humans to solve that ethical quandary and program the automaton to do the same.

There is no "extra liability" with an automaton. Insurance covers hired drivers. Insurance covers autonomous drivers. Insurance companies will jump at the chance to insure the group with a safer driving record.

"An automaton killed my baby!" How is that any different from "a Disney bus driver killed my baby!"? Either way, Disney pays. And by "Disney" I mean "Disney's insurance in both cases." Do you think judges and juries are going to award some super duper extra damages for manslaughter because it was a machine? Do they do that now if someone is killed on a coaster because of operator error v. mechanical failure?

Why,?

Because the legal system has not yet defined liability in the case of autonomous vehicles.

Really? If Kia's car explodes because Kia made the car wrong, then wouldn't that be the same if Kia's autonomy program malfunctioned and drove the passengers off a cliff? Kia pays in both cases.

If there are two Kias, one with a human driver and one with an autonomous driver and they both rear-end another vehicle, isn't that going to be exactly the same with regard to liability, which is that the auto insurance on both cars pays for damages and injury?

 

[21stamps]

The technology for autonomous vehicles is advancing faster than most people realize. Higher end passenger cars already have many seni-autonomous features but the bigger advances will come first in the commercial sector because of the economics of investment returns, the shortage of commercial drivers, and yes the liability issues. We will soon reach the point where computers are safer drivers than humans.

I posted a thread about that last month.

The technology is already there.. what car companies are doing now is easing people into it. Like you said, Most new cars already have autonomous features, some more than others. But out of the percentage who own them- how many people actually use the cruise control, parking assist, etc? I think some features are used more than others, and then there's a psychological factor where people feel "more safe" having these options.
But we have a long way to go before accepting fully autonomous vehicles.. my opinion.

 

[21stamps]

I'm sure Tesla would love to be a part of it... I'm also sure that GM would at least equally like them kept out, especially since they must want THEIR tech in this area showed off...

IMHO, the only way for Tesla to be part of the party is either 1) GM ends their sponsorship or 2) GM & Tesla form a business partnership.

Yep. And I don't see a partnership happening

 

[GoofGoof]

There is no "extra liability" with an automaton. Insurance covers hired drivers. Insurance covers autonomous drivers. Insurance companies will jump at the chance to insure the group with a safer driving record.

"An automaton killed my baby!" How is that any different from "a Disney bus driver killed my baby!"? Either way, Disney pays. And by "Disney" I mean "Disney's insurance in both cases." Do you think judges and juries are going to award some super duper extra damages for manslaughter because it was a machine? Do they do that now if someone is killed on a coaster because of operator error v. mechanical failure?

Actually I think there could be a different level of liability. This is a new and unproven technology. If Disney relies on it and it fails they may have a higher level of liability than if a bus driver gets in an accident. You may not feel that is fair, but there is a good chance that's how it would be viewed in a court of law. If Disney hires a licensed and trained bus driver and they make a human error Disney will be held liable for actual damages (medical, lost wages, damages to vehicles, etc...) but there is no gross negligence on the company's part and so there should be no punitive damages. If a good lawyer convinces the jury that Disney should have known about or anticipated a system failure on an automated shuttle or that they were not properly tested or programmed for a given scenario they could push for gross negligence and huge punitive damages. Insurance won't cover punitive damages.

Again, I'm talking about having automated shuttles on highways with other cars traveling in excess of 50 MPH. There's way too many variables and things that could go wrong to cover all the scenarios with programming. Slow moving shuttles on dedicated roadways eliminate most of those variables, but it limits the uses somewhat.

 

[Pam Hates Penguins]

What kind of world is this? We are not in 2030!

 

[MisterPenguin]

Actually I think there could be a different level of liability. This is a new and unproven technology. If Disney relies on it and it fails they may have a higher level of liability than if a bus driver gets in an accident. You may not feel that is fair, but there is a good chance that's how it would be viewed in a court of law. If Disney hires a licensed and trained bus driver and they make a human error Disney will be held liable for actual damages (medical, lost wages, damages to vehicles, etc...) but there is no gross negligence on the company's part and so there should be no punitive damages. If a good lawyer convinces the jury that Disney should have known about or anticipated a system failure on an automated shuttle or that they were not properly tested or programmed for a given scenario they could push for gross negligence and huge punitive damages. Insurance won't cover punitive damages.

Again, I'm talking about having automated shuttles on highways with other cars traveling in excess of 50 MPH. There's way too many variables and things that could go wrong to cover all the scenarios with programming. Slow moving shuttles on dedicated roadways eliminate most of those variables, but it limits the uses somewhat.

 

[ford91exploder]

Here is a primer on product liability for the non professional

https://www.travelers.com/resources...gid=10907348&dclid=CLeqoOC13dMCFce2swodZggAVA

 

[the.dreamfinder]

Putting the short to medium term challenges of autonomous driving aside and assuming they will be overcome, let's look at the urban planning and congestion issues autonomous shuttles would bring to a medium to large city, like WDW. Published the same week as Elon Musk released his vision for a network underground high speed skateboards for automobiles, The Verge delves into the negative effects of autonomous cars on congestion.
https://www.theverge.com/2017/5/6/1...company-uber-flying-car-public-transportation

 

[Kman101]

You do know that there are autonomous vehicles on the roads today out in the real world, right? It won't take ten years.

But they aren't everywhere ...

 

[21stamps]

But they aren't everywhere ...

And they still won't be in 10 years from now. The market will grow, but I'll predict 20 years..or more..before we see a majority of fully autonomous vehicles sharing the roads with a minority non autonomous.

 

[GoofGoof]

And they still won't be in 10 years from now. The market will grow, but I'll predict 20 years..or more..before we see a majority of fully autonomous vehicles sharing the roads with a minority non autonomous.

This is the point I'm trying to make. The technology is evolving but the article that started this discussion is stating that WDW will roll out test vehicles this year starting with CM transport with guests using them starting next year. There's zero chance what they roll out is fully autonomous shuttles that run at full speed on roads with regular cars. That may come some time in the future...10 or 20 years or longer who knows, but it's not ready now and Disney is very unlikely to want to take the risks associated with being a first mover and use its resort guests as test subjects.

 

[MisterPenguin]

Putting the short to medium term challenges of autonomous driving aside and assuming they will be overcome, let's look at the urban planning and congestion issues autonomous shuttles would bring to a medium to large city, like WDW. Published the same week as Elon Musk released his vision for a network underground high speed skateboards for automobiles, The Verge delves into the negative effects of autonomous cars on congestion.
https://www.theverge.com/2017/5/6/1...company-uber-flying-car-public-transportation

That's not the takeaway I get from that article. It's a rant about flying cars and previous bad mass-transportation mistakes. It doesn't touch any studies of what autonomous cars can do.

 

[the.dreamfinder]

That's not the takeaway I get from that article. It's a rant about flying cars and previous bad mass-transportation mistakes. It doesn't touch any studies of what autonomous cars can do.

So having a city full of autonomous vehicles driving around waiting for customers isn't a mistake in the making? Because cities need to plan for higher density transportation solutions such as light and heavy rail, bike lanes and intermodal transportation facilities to connect everything together.

 

[21stamps]

This is the point I'm trying to make. The technology is evolving but the article that started this discussion is stating that WDW will roll out test vehicles this year starting with CM transport with guests using them starting next year. There's zero chance what they roll out is fully autonomous shuttles that run at full speed on roads with regular cars. That may come some time in the future...10 or 20 years or longer who knows, but it's not ready now and Disney is very unlikely to want to take the risks associated with being a first mover and use its resort guests as test subjects.

Agree. I think it would be neat if GM would incorporate the technology into some sort of Epcot based attraction, maybe showcase it somewhere within Test Track.

So having a city full of autonomous vehicles driving around the city waiting for customers isn't a mistake in the making? Because cities need to plan for higher density transportation solutions such as light and heavy rail, bike lanes and intermodal transportation facilities to connect everything together.

I don't think that will ever be the reality in our lifetime, so I'm not too worried about it now.

 

[GoofGoof]

So having a city full of autonomous vehicles driving around the city waiting for customers isn't a mistake in the making? Because cities need to plan for higher density transportation solutions such as light and heavy rail, bike lanes and intermodal transportation facilities to connect everything together.

This is true. More efficient mass transit in cities is the better answer. I think autonomous vehicles have some potential, but I feel like you still need a driver. It's kinda like an airplane now. For most of the flight the plane flies itself, but they don't just let the plane take off without pilots. Adding more and more technology to cars to make them more automated and remove more of the possible human errors is a positive thing. I think fully autonomous vehicles really only work in limited, controlled situations or if all vehicles are driverless. If everything was driverless than the vehicles could easily predict each other's moves because they would all be operating on the same set of rules. The problem with human drivers is they don't always follow the rules or even basic logic. It's much harder to predict.

 

[21stamps]

This is true. More efficient mass transit in cities is the better answer. I think autonomous vehicles have some potential, but I feel like you still need a driver. It's kinda like an airplane now. For most of the flight the plane flies itself, but they don't just let the plane take off without pilots. Adding more and more technology to cars to make them more automated and remove more of the possible human errors is a positive thing. I think fully autonomous vehicles really only work in limited, controlled situations or if all vehicles are driverless. If everything was driverless than the vehicles could easily predict each other's moves because they would all be operating on the same set of rules. The problem with human drivers is they don't always follow the rules or even basic logic. It's much harder to predict.

Yep. And our roads don't even allow for it right now. Read your owner's manual about Lane Assist (or any similarly named technology) it "may not work on winding roads, roads with faint lines, roads with bumps,...." etc etc etc.
even the parking features are the same. The car is looking for lines, I've only used my parking twice successfully, and only 3 times total- twice for perpendicular and once for parallel, just to test them,( I really don't have a desire to use that particular feature). The parallel was fine. Scary, very scary lol, but fine.
The perpendicular wasn't a success the first time- because I was close to a poll on the drivers side, and the line was worn away.. the car disengaged, but a human could have easily done it. The next time, with clear lines, the tech worked perfect.

There are so many variables to relying on cameras and sensors. Same with braking, what if you could swerve instead of brake? What if you see someone in your rear view mirror? ...and so on.

The tech is great, and comforting to have. The more autonomous features then the more safe our roads are becoming. But drivers will be on their proper seats for years to come.

 

[MisterPenguin]

So having a city full of autonomous vehicles driving around waiting for customers isn't a mistake in the making? Because cities need to plan for higher density transportation solutions such as light and heavy rail, bike lanes and intermodal transportation facilities to connect everything together.

We already have cities full of vehicles driving around waiting for customers already with human drivers. So, autonomous ones could hardly make it worse.

Besides, autonomous service vehicles don't have to be on the hunt like taxis, they can find low traffic areas to hang out. A city with cameras pointing at roadside parking can alert autos where to park. Services which provide a fleet of auto-taxis can have a garage or parking lot where the autos wait packed in tight (imagine a whole array of tightly packed cars moving slowly forward at the same time making room for new cars pulling in behind). Autonomous fleets will have algorithms predicting where the heaviest use will be.

And none of this should replace other options: communal bike, rail, gondolas, etc... The self-powered ones will be automated... like the monorail.