Even when the driver is fully responsible, the assistance software must work properly in all situations. And it must be tested fully.
In case the software makes severe mistakes surprisingly, normal drivers maybe don’t have a chance to regain control. Normal drivers are not like educated test drivers.
My morality says both are accountable. The driver, and Tesla. Tesla for damage caused by their system, and the driver for and if he does not retake control of the vehicle given the chance.
But does the driver have a reasonable chance with adequate timeframe to regain control?
Like what happened with Boeing 737 Max MCAS incident, Boeing expects the pilot to disengage the trim motor in mere 4 seconds, which accoriding to a pilot “a lot to ask in an overwheming situation” or something similar.
Normal people in soon-to-crash situation are likely to freeze for a second or two, and the fear kicks up. How the driver reacts next is hard to predict. Yet, at the speed most US drivers love to go (I saw 70+ mph on freeway is the norm), the time avalible for them to make an well thought out decision I guess is quite short.
In my head, the reason is not specifically to punish the driver, but to make drivers always be aware and ready to take control again. Yes 100 ppl will have 1000 different ways to react to such a software error, but you need ppl to pay attention, and in law the only way is to use punishment. Obviously this needs to be well calculated but either you have multiple lines of defense (the software, the driver, maybe even additional safety features) or you have to remove the autonomous system.
People are naturally going to pay less attention the more cars drive for them. You can’t partially automate steering. Driver assisted steering is as close as it can be before the liability needs to fall on Tesla and other software manufacturers. A car isn’t a plane. The driver needs to be in control when split second decisions happen, like a child running after a ball.
If I’m paying for an autopilot, I’m not the pilot. I.e., the driver. The car is. And Tesla’s marketing bullshit and lawyers are going to fail here. This does not fall under puffery. It’s false advertising that’s causing consumers to place undue trust in a product. And the insurance industry is quite concerned just where the liability falls in all of this as well. And as they’re the ones currently having to pay out claims when Tesla wins, they have a vested interest seeing that Tesla doesn’t.
It doesn’t matter for practical purposes you can’t make people pay attention as if driving without the actual engagement of driving. There is going to be a delay in taking over and in a lot of cases it wont matter by the time the human is effectively in control.
Imagine you are going along a straight road, not too much traffic, the speed limit is high and you are enjoying it. Suddenly your assistant software decides to turn your steering wheel hard to the left.
You will have no chance.
What have you done wrong? What is it what you are accountable for?
The article keeps calling it “Autopilot”, which is different from “Full Self Driving”.
If they are correct, then it’s all on the driver. Autopilot is just a nicer adaptive cruise control, and should be treated as such. Many cars have them, even non-smart vehicles. Even my seven year old Subaru had similar (much dumber but similar)
That being said, people seem to confuse the names of these different functionalities all the time, including throughout this thread. However, even if they were confused and meant FSD, my car has feedback to require your hands in the wheel, so I don’t understand how you can claim ignorance
Even when the driver is fully responsible, the assistance software must work properly in all situations. And it must be tested fully.
In case the software makes severe mistakes surprisingly, normal drivers maybe don’t have a chance to regain control. Normal drivers are not like educated test drivers.
My morality says both are accountable. The driver, and Tesla. Tesla for damage caused by their system, and the driver for and if he does not retake control of the vehicle given the chance.
But does the driver have a reasonable chance with adequate timeframe to regain control?
Like what happened with Boeing 737 Max MCAS incident, Boeing expects the pilot to disengage the trim motor in mere 4 seconds, which accoriding to a pilot “a lot to ask in an overwheming situation” or something similar.
Normal people in soon-to-crash situation are likely to freeze for a second or two, and the fear kicks up. How the driver reacts next is hard to predict. Yet, at the speed most US drivers love to go (I saw 70+ mph on freeway is the norm), the time avalible for them to make an well thought out decision I guess is quite short.
You made me think about this for a second.
In my head, the reason is not specifically to punish the driver, but to make drivers always be aware and ready to take control again. Yes 100 ppl will have 1000 different ways to react to such a software error, but you need ppl to pay attention, and in law the only way is to use punishment. Obviously this needs to be well calculated but either you have multiple lines of defense (the software, the driver, maybe even additional safety features) or you have to remove the autonomous system.
People are naturally going to pay less attention the more cars drive for them. You can’t partially automate steering. Driver assisted steering is as close as it can be before the liability needs to fall on Tesla and other software manufacturers. A car isn’t a plane. The driver needs to be in control when split second decisions happen, like a child running after a ball.
If I’m paying for an autopilot, I’m not the pilot. I.e., the driver. The car is. And Tesla’s marketing bullshit and lawyers are going to fail here. This does not fall under puffery. It’s false advertising that’s causing consumers to place undue trust in a product. And the insurance industry is quite concerned just where the liability falls in all of this as well. And as they’re the ones currently having to pay out claims when Tesla wins, they have a vested interest seeing that Tesla doesn’t.
It doesn’t matter for practical purposes you can’t make people pay attention as if driving without the actual engagement of driving. There is going to be a delay in taking over and in a lot of cases it wont matter by the time the human is effectively in control.
Imagine you are going along a straight road, not too much traffic, the speed limit is high and you are enjoying it. Suddenly your assistant software decides to turn your steering wheel hard to the left.
You will have no chance.
What have you done wrong? What is it what you are accountable for?
For mine
So did the car think there was an impending collision? That should be obvious in the logs and the only reason for sudden maneuvers
Cars do not think LOL
The article keeps calling it “Autopilot”, which is different from “Full Self Driving”.
If they are correct, then it’s all on the driver. Autopilot is just a nicer adaptive cruise control, and should be treated as such. Many cars have them, even non-smart vehicles. Even my seven year old Subaru had similar (much dumber but similar)
That being said, people seem to confuse the names of these different functionalities all the time, including throughout this thread. However, even if they were confused and meant FSD, my car has feedback to require your hands in the wheel, so I don’t understand how you can claim ignorance
No. That difference is meaningless, since both softwares provide autonomy level 2. The responsibilities are exactly the same.