Mercedes solves two problems at once R/robotlaw

[Tcoat]

That ain't nothin'
Wait until the Mach E gets self driving abilities

[p1l0t]

Originally Posted by extrashaky

I anticipate that self-driving cars will turn out to be far less popular than people expect. Here's why:

The AI will be programmed to leave a safe stopping distance between it and the car in front of it. It will be programmed not to drive assertively, but to give way to more assertive drivers. It will automatically take evasive action when someone changes lanes or pulls out into traffic.

This type of programming would actually reward the more assertive drivers in other cars. People in control of their own cars would be able to zip around the self-driving cars, especially once their driving style is known and so utterly predictable that everyone on the road will be able to spot the cars in automatic mode.

Now consider the people sitting in the self-driving cars. I've driven a lot of places, and one common thread is that drivers really don't like other people "beating" them at traffic. The guy sitting there letting the car drive will get irritated that the AI keeps letting people break in line, and he'll take it out of auto to drive it himself. Once you have enough people doing that, any collective benefit of self-driving cars is eliminated.

Whenever you see a self-driving car in a sci-fi movie, the passenger is always sitting there reading a newspaper as if he's on a train or subway, completely ignoring the outside world. But a self-driving car is not like a train. When you're on a train, you're not watching other more assertive trains pass you. And even if you did, it's completely out of your control, unlike an automatic car that you could switch to manual mode to do traffic battle.

The only place the self-driving mode will be attractive to non-commercial drivers is on long drives on sparsely populated roads. Most people rarely make long drives on sparsely populated roads. The overwhelming majority of time spent in cars is within a few miles of the driver's home, and a large percentage of the long distance driving is on fairly crowded interstates, where the cars would be taken out of automatic to negotiate traffic.

And once a few affordable cars have an automatic mode and people realize how little they will use it, they'll forego the cost of it and buy the cheaper version without a self-driving mode.

Most of what you say is probably true, except the few minutes lost to the aggressive drivers is made up by the fact that you can be doing something else with your time. Often I have Uber from one airport to another because there are no crew cars or rental cars available. Most* of those drivers won't do more than 5 over and tend to stay out of the passing lane (as they should). I know I could do what they do in an hour and ten, in fifty minutes. It doesn't bother though because while they are driving I'm filing flight plans, doing weight and balance calculations, figuring which airport on my route has the cheapest fuel, checking the weather for departure/enroute/arrival, etc.. so even though I have less time before wheels up it actually saves me time overall. It costs the company more though generally.

Sent from my SM-G975U using Tapatalk

[dowroa]

Originally Posted by extrashaky

I anticipate that self-driving cars will turn out to be far less popular than people expect. Here's why:

The AI will be programmed to leave a safe stopping distance between it and the car in front of it. It will be programmed not to drive assertively, but to give way to more assertive drivers. It will automatically take evasive action when someone changes lanes or pulls out into traffic.

This type of programming would actually reward the more assertive drivers in other cars. People in control of their own cars would be able to zip around the self-driving cars, especially once their driving style is known and so utterly predictable that everyone on the road will be able to spot the cars in automatic mode.

Now consider the people sitting in the self-driving cars. I've driven a lot of places, and one common thread is that drivers really don't like other people "beating" them at traffic. The guy sitting there letting the car drive will get irritated that the AI keeps letting people break in line, and he'll take it out of auto to drive it himself. Once you have enough people doing that, any collective benefit of self-driving cars is eliminated.

Whenever you see a self-driving car in a sci-fi movie, the passenger is always sitting there reading a newspaper as if he's on a train or subway, completely ignoring the outside world. But a self-driving car is not like a train. When you're on a train, you're not watching other more assertive trains pass you. And even if you did, it's completely out of your control, unlike an automatic car that you could switch to manual mode to do traffic battle.

The only place the self-driving mode will be attractive to non-commercial drivers is on long drives on sparsely populated roads. Most people rarely make long drives on sparsely populated roads. The overwhelming majority of time spent in cars is within a few miles of the driver's home, and a large percentage of the long distance driving is on fairly crowded interstates, where the cars would be taken out of automatic to negotiate traffic.

And once a few affordable cars have an automatic mode and people realize how little they will use it, they'll forego the cost of it and buy the cheaper version without a self-driving mode.

If this world comes to be, you think they will allow people to drive their own cars on these roads with the "more efficient, safer, paying" self-driving cars?

What about, if you wanted to enter $city_name_here, you could only do so in a rented self-driving, zero-emission, city approved and allotted conveyance that paid tithes to the city?

That form of payment is already making its way into cities with the scooter laws.

[soundman98]

Originally Posted by Spuds

If it had a choice to hit a bunch of old people or a gaggle of children, would it hit the children because old people are more likely to be customers?

your question is null, as the children would inherit the money from the old's to spend on a mercedes. clearly the only option here is to take out the old's.

Originally Posted by extrashaky

I anticipate that self-driving cars will turn out to be far less popular than people expect. Here's why:

The AI will be programmed to leave a safe stopping distance between it and the car in front of it. It will be programmed not to drive assertively, but to give way to more assertive drivers. It will automatically take evasive action when someone changes lanes or pulls out into traffic.

This type of programming would actually reward the more assertive drivers in other cars. People in control of their own cars would be able to zip around the self-driving cars, especially once their driving style is known and so utterly predictable that everyone on the road will be able to spot the cars in automatic mode.

Now consider the people sitting in the self-driving cars. I've driven a lot of places, and one common thread is that drivers really don't like other people "beating" them at traffic. The guy sitting there letting the car drive will get irritated that the AI keeps letting people break in line, and he'll take it out of auto to drive it himself. Once you have enough people doing that, any collective benefit of self-driving cars is eliminated.

Whenever you see a self-driving car in a sci-fi movie, the passenger is always sitting there reading a newspaper as if he's on a train or subway, completely ignoring the outside world. But a self-driving car is not like a train. When you're on a train, you're not watching other more assertive trains pass you. And even if you did, it's completely out of your control, unlike an automatic car that you could switch to manual mode to do traffic battle.

The only place the self-driving mode will be attractive to non-commercial drivers is on long drives on sparsely populated roads. Most people rarely make long drives on sparsely populated roads. The overwhelming majority of time spent in cars is within a few miles of the driver's home, and a large percentage of the long distance driving is on fairly crowded interstates, where the cars would be taken out of automatic to negotiate traffic.

And once a few affordable cars have an automatic mode and people realize how little they will use it, they'll forego the cost of it and buy the cheaper version without a self-driving mode.

there was an article about this from google's self driving program quite a while back(i want to say pre-tesla), and how they were starting to program their cars to be more agressive like regular drivers.

Originally Posted by dowroa

What about, if you wanted to enter $city_name_here

found the programmer!

[extrashaky]

Originally Posted by p1l0t

Most of what you say is probably true, except the few minutes lost to the aggressive drivers is made up by the fact that you can be doing something else with your time.

You can be, but a large percentage (if not most) people will not do something else because they'll be distracted by what's going on outside the car.

Originally Posted by p1l0t

Often I have Uber from one airport to another because there are no crew cars or rental cars available.

Will the Uber driver allow you to take over and drive? I'm guessing not. Therefore Uber is irrelevant.

In an Uber you can put the outside world out of your mind and do other work because you have relinquished control. You don't completely relinquish control to an automated car. You have the option of taking back the wheel.

In fact, with the current fleet of self-driving cars I believe you're required to pay attention and be ready to take over. You're not required to be ready to take the wheel from the Uber driver.

Originally Posted by dowroa

If this world comes to be, you think they will allow people to drive their own cars on these roads with the "more efficient, safer, paying" self-driving cars?

Yes, absolutely, for two reasons.

One, people vote. Once people get a real taste of self-driving cars, even the ones who think it's a good idea now will vote against any lawmaker attempting to make them mandatory.

Two, your assumption that they are "more efficient, safer" is a losing battle with respect to persuading the voters to accept them.

I already covered the efficiency aspect a little, but to elaborate, traffic jams are exacerbated by people doing stupid shit. It's true that if everyone drove at the proper speed with the proper distance between vehicles, traffic would be more efficient and jams would be reduced. It's also true that people don't drive that way because they are afraid other people will cut in line in front of them, so they follow too closely, causing them to have to stop abruptly, which creates caterpillar-style waves in the traffic that slow it down. And it's true that AI cars could coordinate to achieve what humans can't.

But that fails to acknowledge the humans in the car who will take over when they see the AI giving way to more assertive human drivers or who just become fed up with watching the AI pass up opportunities to get ahead. To them, the other guy is cheating, the AI is letting him cheat, and as a result the AI is losing while someone else is winning. In a laboratory setting the AI-controlled cars will almost certainly improve traffic flow. In the real world I anticipate that they'll see no improvement at all or even worse performance when mixed with human drivers.

As for whether they're safer, I doubt you would see a decrease in accident rates in the real world. When seat belts were introduced, the death rate among drivers remained the same. You were more likely to survive a crash, but you were also more likely to be in a crash in the first place, because seat belts made people feel it was safer to drive faster and more recklessly. When those same people drive around self-driving cars they know will give way, they'll take more chances.

Meanwhile, even if the actual stats were in favor of the AI, voters who hate the idea of not being in control would not be convinced by that and would instead be more persuaded by the inevitable anecdotal evidence about crashes involving self-driving cars. Man vs. evil machines is something that has been ingrained in our collective psyche since before John Henry defeated the steam engine. We're not about to let Skynet take over our cars now.

Originally Posted by dowroa

What about, if you wanted to enter $city_name_here, you could only do so in a rented self-driving, zero-emission, city approved and allotted conveyance that paid tithes to the city?

A few misguided city councils will probably try that. But then most will abandon the idea once the tax revenue declines from decreased business in the exclusion zones. Why would I drive to the city, find parking and pay for a shitty robot car to take me the rest of the way if I can just do all my business in the suburbs or over the internet instead? Companies will recognize the added stress of getting into these zones and will move outside them, for the benefit of both their customers and their employees.

There will likely be a few exclusion zones in northeast cities in locations where most people already use public transportation to get to work and won't be affected by the restrictions. The self-driving car lobby will point to those and call them successes, even though self-driving cars will have nothing to do with it.

Then they'll try it in a place like Atlanta, thinking it will solve the ridiculous traffic there, and watch it fail spectacularly.

Originally Posted by soundman98

there was an article about this from google's self driving program quite a while back(i want to say pre-tesla), and how they were starting to program their cars to be more agressive like regular drivers.

It won't actually happen. Or rather, if it does happen, it will get rolled back so fast that the end result will be the same. The reason is that you'd be mixing two different kinds of legal liability.

Suppose one of these cars with the more assertive programming gets into an accident. Suppose it's not even the AI's fault, but the human driver in the other vehicle caused the accident.

Under our personal liability statutes for motor vehicles, the human driver's insurance company would clearly be liable.

But under product liability statutes, a good lawyer would argue that Google's AI was intentionally made less safe, and therefore Google was responsible for the accident.

Google might say, "But we designed our AI to drive like a human! Human rules should apply!"

And the lawyer will say, "But it's NOT human. It's an AI, which could have avoided this accident if you hadn't dumbed it down."

And the jury will say, "Human driver wins! One billion dollars!"

We'll never get to that point, because the lawyers at Google, Tesla, Ford and every other company developing AI know better and will nuke that idea before it ever gets a firm foothold.

[Clutch Dog]

Originally Posted by dowroa

If this world comes to be, you think they will allow people to drive their own cars on these roads with the "more efficient, safer, paying" self-driving cars?

What about, if you wanted to enter $city_name_here, you could only do so in a rented self-driving, zero-emission, city approved and allotted conveyance that paid tithes to the city?

That form of payment is already making its way into cities with the scooter laws.

im actually writing up the concept for this, with full integration into say the harbor and aviation systems. im basing it off of san diego as its overly complex but i think it could work

the issue is getting over "i dont feel like im free cause i dont drive my own cars" and the odd "handicap" problems that never seem to be accounted for

[p1l0t]

Originally Posted by extrashaky

You can be, but a large percentage (if not most) people will not do something else because they'll be distracted by what's going on outside the car.



Will the Uber driver allow you to take over and drive? I'm guessing not. Therefore Uber is irrelevant.

In an Uber you can put the outside world out of your mind and do other work because you have relinquished control. You don't completely relinquish control to an automated car. You have the option of taking back the wheel.

In fact, with the current fleet of self-driving cars I believe you're required to pay attention and be ready to take over. You're not required to be ready to take the wheel from the Uber driver.



Yes, absolutely, for two reasons.

One, people vote. Once people get a real taste of self-driving cars, even the ones who think it's a good idea now will vote against any lawmaker attempting to make them mandatory.

Two, your assumption that they are "more efficient, safer" is a losing battle with respect to persuading the voters to accept them.

I already covered the efficiency aspect a little, but to elaborate, traffic jams are exacerbated by people doing stupid shit. It's true that if everyone drove at the proper speed with the proper distance between vehicles, traffic would be more efficient and jams would be reduced. It's also true that people don't drive that way because they are afraid other people will cut in line in front of them, so they follow too closely, causing them to have to stop abruptly, which creates caterpillar-style waves in the traffic that slow it down. And it's true that AI cars could coordinate to achieve what humans can't.

But that fails to acknowledge the humans in the car who will take over when they see the AI giving way to more assertive human drivers or who just become fed up with watching the AI pass up opportunities to get ahead. To them, the other guy is cheating, the AI is letting him cheat, and as a result the AI is losing while someone else is winning. In a laboratory setting the AI-controlled cars will almost certainly improve traffic flow. In the real world I anticipate that they'll see no improvement at all or even worse performance when mixed with human drivers.

As for whether they're safer, I doubt you would see a decrease in accident rates in the real world. When seat belts were introduced, the death rate among drivers remained the same. You were more likely to survive a crash, but you were also more likely to be in a crash in the first place, because seat belts made people feel it was safer to drive faster and more recklessly. When those same people drive around self-driving cars they know will give way, they'll take more chances.

Meanwhile, even if the actual stats were in favor of the AI, voters who hate the idea of not being in control would not be convinced by that and would instead be more persuaded by the inevitable anecdotal evidence about crashes involving self-driving cars. Man vs. evil machines is something that has been ingrained in our collective psyche since before John Henry defeated the steam engine. We're not about to let Skynet take over our cars now.



A few misguided city councils will probably try that. But then most will abandon the idea once the tax revenue declines from decreased business in the exclusion zones. Why would I drive to the city, find parking and pay for a shitty robot car to take me the rest of the way if I can just do all my business in the suburbs or over the internet instead? Companies will recognize the added stress of getting into these zones and will move outside them, for the benefit of both their customers and their employees.

There will likely be a few exclusion zones in northeast cities in locations where most people already use public transportation to get to work and won't be affected by the restrictions. The self-driving car lobby will point to those and call them successes, even though self-driving cars will have nothing to do with it.

Then they'll try it in a place like Atlanta, thinking it will solve the ridiculous traffic there, and watch it fail spectacularly.



It won't actually happen. Or rather, if it does happen, it will get rolled back so fast that the end result will be the same. The reason is that you'd be mixing two different kinds of legal liability.

Suppose one of these cars with the more assertive programming gets into an accident. Suppose it's not even the AI's fault, but the human driver in the other vehicle caused the accident.

Under our personal liability statutes for motor vehicles, the human driver's insurance company would clearly be liable.

But under product liability statutes, a good lawyer would argue that Google's AI was intentionally made less safe, and therefore Google was responsible for the accident.

Google might say, "But we designed our AI to drive like a human! Human rules should apply!"

And the lawyer will say, "But it's NOT human. It's an AI, which could have avoided this accident if you hadn't dumbed it down."

And the jury will say, "Human driver wins! One billion dollars!"

We'll never get to that point, because the lawyers at Google, Tesla, Ford and every other company developing AI know better and will nuke that idea before it ever gets a firm foothold.

I was just making the point that if you can do something else then you can probably make up the lost time and then some. That's why I used Uber as an example because even though most drivers are slow, I can do most of the stuff I would be doing in the pilot office anyway. So yeah the trip takes an extra 10 or 20 minutes even on a long drive but I can bang out 30-45min worth of planning. So I don't even need to use the pilot office except to print out the results and then I just go straight to the plane, figure out if my S.I.C. needs anything, and we blast off. It's like a net gain of 20 minutes really if you don't waste the time worrying about who is passing you. This would only work (safely) with full automation though. Otherwise you need a driver that's not you.

Sent from my SM-G975U using Tapatalk

[Tcoat]

Originally Posted by extrashaky

In an Uber you can put the outside world out of your mind and do other work because you have relinquished control. You don't completely relinquish control to an automated car. You have the option of taking back the wheel.

This is EXACTLY what most drivers on the road will do. Once they don't have to pay attention they won't. Hell they don't now when the car is not self driving so why would they all of a sudden want to when it is?

[extrashaky]

Originally Posted by Tcoat

This is EXACTLY what most drivers on the road will do. Once they don't have to pay attention they won't. Hell they don't now when the car is not self driving so why would they all of a sudden want to when it is?

Some will do that. I don't believe most will.

People are not consistently attentive or inattentive. Drivers are in active competition with the world around them, then suddenly lose interest when a text comes in, then go into combat mode when someone passes. I don't know how many times I've been behind some idiot who was speeding up and slowing down because he/she was on the phone and not paying attention to driving, only to have that same driver speed up to try to block me when I try to go around.