Two men stood in front of the autonomous vehicle, operated by ride-hailing company Waymo, and literally tipped a fedora at her while she told them to move out of the way.
These cars need to have a panic button that allows a remote operator to talk to the passengers, assess the situation, alert police and override the auto driving to get them out of bad situations. Same as an emergency call button on an elevator basically. I dont understand these cars to have any feature like that so far, and I’m assuming this woman would have used it if one was available, so please correct me if I’m wrong.
These cars are likely going to turn into hijack machines if they’re programmed for “maximum safety” in situations where, realistically, breaking every traffic law, hitting a pedestrian or causing damage to the vehicle through dangerous terrain may be the only way out with a living passenger. The second it begins to percolate among criminals that these things are super easy to stop at the perfect location of your choosing like this, they are going to become a massive target.
Or they turn into a hearse if the passenger has a medical emergency and the car doesn’t redirect while the passenger is incapacitated. They might be coherent enough to press a button, but not to open their phone, navigate the app, call for help or redirect the car to a hospital…
But that of course requires labor so it will not happen until legally mandated after a minimum threshold of people die.
“The men came over to the car again and stood in front of it for a few minutes. Finally when they left, the car was still stalled but I clicked the ‘in car support’ on the screen and they seemed to be aware of the issue,” Amina said. “They asked if I was OK and the car began to drive towards my location. They asked if I needed police support and I said no.”
When she was almost to her destination, Waymo support called her again to ask if she was ok, she said. “I assured him that I was fine and he told me I would be given a free ride after,” she said. “After many hours I was called one last time by their support team. They asked if I was OK and told me that they have 24/7 support available. They also said I would get the next ride or next two rides (uncertain) free.”
“In an instance like this, our riders have 24/7 access to Rider Support agents who will help them navigate the situation in real time and coordinate closely with law enforcement officers to provide further assistance as needed,” a spokesperson for Waymo told 404 Media in an email. “While these sorts of events are exceedingly rare among the 100,000 trips we serve a week across Los Angeles, San Francisco, and Phoenix, we take them very seriously. We continuously look for ways to improve rider experience and remain committed to improving road safety and mobility in the cities where we operate.”
Agreed, but to play devil’s advocate, the support wasn’t branded as such and customers could’ve not reported out of shame, which wouldn’t happen if they knew they could do that at the beginning before it became anything substantial.
They have customer support that provides words of platitude, an ineffective police call with a 15minute response time, and no control over the situation. She got lucky this time, but my point remains standing.
I must be tired right now but I don’t see how a remote operator could have driven better in this situation.
You can’t get away from someone blocking your car in traffic without risk.of hitting them or other people or vehicles.
You probably meant they ought to drive away regardless of what they hit, if it helps the passenger escape a.dire.situation? But I have to wonder if a remote operator would agree to be put on the spot like that.
Yea I’m not too keen on giving authorization to hit pedestrians. If I feel threatened in my car, I am not allowed to run over the person so why should a driverless car gain that right? And if the panic button is going to call the police, how is that any different from the passenger using their phone to contact police? Seems like extra steps of middlemen and confusion when the passenger could just call once they feel the need.
I could defintely see a case for some extra safety features that help keep the doors locked and shut, maybe thicker windows too if needed to prevent robberies/assaults.
If I feel threatened in my car, I am not allowed to run over the person
You are not allowed to run people over merely because you feel threatened.
You are allowed to use deadly force, in the USA when you reasonably believe that it is necessary to prevent someone from unlawfully killing, causing serious physical injury, or committing a short list of violent felonies. The harassment described in the article probably does not rise to that level, though an ambitious lawyer might try to describe intentionally causing the car to stop as carjacking or kidnapping.
It’s likely the harassers can be prosecuted for false imprisonment, a misdemeanor. It is illegal to use deadly force such as hitting people with cars to prevent/terminate a misdemeanor.
Is there any law in any state that would allow you to kill a 3rd party to escape being killed yourself? (If there were, I’d probably opt for not living in that state)
What do you mean by “allow you to kill a 3rd party”?
Like if rioters are breaking into your window and start trying to pull you out through it, then you floor it and kill someone else in the crowd who wasn’t actively breaking into your car?
This is something that’s going to vary from state to state, but ultimately it will be a case by case decision where a jury will decide if the use of deadly force was reasonable.
You will be judged based on other’s perception of the events, not based solely whether you yourself thought you were in danger or not.
So, someone trying to “drive slowly” through a group of protesters would probably be found at fault, while a car that was stuck trying to wait patiently suddenly having a Molotov cocktail thrown on it would be judged differently. Even then they will need to consider whether you could have just gotten out of your car and run.
I vaguely remember reading in my criminal law textbook, years back, that murder is one of the few exceptions to the doctrine of necessity (this would have been in the context of US law), so I don’t think that it’s ever legally-permissible to intentionally kill some random person to save yourself. IIRC the rationale was that it prevents thing like terrorist groups from coercing someone to do actions for them by threatening someone else.
That being said, there are obviously points where people are forced to take actions where either one group of people is going to die or another; in ethics, the trolley problem is a well-known example. For a maybe-less-artificial problem, closing hatches in a ship where not everyone is out of a compartment to prevent the ship from going down, say. I don’t know how law applies in the situation of weighing lives; my assumption is that it doesn’t mandate inaction.
The “hitting pedestrians” is an extreme hypothetical, and not one you should particularly get hung up on. But it is one that still has to be considered. Passive security measures only go so far for the passenger.
Realistically, a car can get out of a vast majority of situations evasively without hitting hostile pedestrians, such as reversing rapidly and then turning around or driving in an opposite travel lane to bypass the blockage. Or hopping a curb and using a sidewalk if it is not occupied (or just blasting the shit out of the horn if it is occupied). These are all things that waymo’s auto mode cannot and will not do, because it doesn’t have the reasoning to understand when such measures are necessary.
If you legitimately believe your life is in danger, you have the right to escape or defend yourself, even if that means running someone over. This has happened in multiple countries with similar outcomes.
And if the panic button is going to call the police, how is that any different from the passenger using their phone to contact police? Seems like extra steps of middlemen and confusion when the passenger could just call once they feel the need.
Think of it as a backup for the phone in the case where, say, there’s an adult and a kid in the car, the kid has no phone of their own, and the adult loses consciousness with their phone locked. Or the car is being actively jostled by a group of people (say it drove into the middle of an embryonic riot), causing the passenger to drop their phone, whereupon it slides under the seat. Or the phone just runs out of charge or doesn’t survive getting dropped into the passenger’s triple-extra-large fast-food coffee. It won’t be needed 99% of the time, but the other 1% might save someone’s life, and (presuming the car already has a cell modem it in) the cost of adding the feature should be minimal.
If a man jumps out in front of my car in traffic and points a pistol at me after I stop. I am going around or thru him and there is no other option. Anyone else trying to stop me even without visible weapons is going to get evasive maneuvers to protect myself because I am not dealing with that bullshit. That includes weaving far outside my travel lane or going over a sidewalk. That is self defense and a split second decision that any driver may have to make. Waymo prioritizes all outside obstacle avoidance which means it doesn’t even want to leave it’s set travel lane, which makes them trivial to stop like this with no recourse.
The point I am making is that self driving has a really hard time interpreting traffic edge cases or passenger emergencies like this. A remote operator could make the decision to drive over curbs and other lanes, if free, to save the passenger, and realistically should avoid hitting pedestrians too… but in the case of an armed attacker - well, yknow. Like force for like force.
Calling police would only be an auxiliary function to report the video evidence. They cannot be depended on to respond in time to actually make a difference.
Would a remote operator interpret things accurately in 10 seconds or less, or be a job anyone would even want? How does the liability chain of command work? Who knows. But the current system makes no decision at all, and that is unacceptable. And the medical point still stands too, a remote operator could immediately reroute the vehicle to a hospital and alert the medical staff. A panic button is absolutely needed.
It worked, only because these men were only being creepy sexist pieces of shit and didn’t have worse intentions. Customer support according to this article has no control over the vehicle other than restarting the auto driving routines to make the car move again.
Well and the draw of these tiny driverless train like objects kinda goes out the window when you have to staff anything at all to monitor and control them.
These cars need to have a panic button that allows a remote operator to talk to the passengers, assess the situation, alert police and override the auto driving to get them out of bad situations. Same as an emergency call button on an elevator basically. I dont understand these cars to have any feature like that so far, and I’m assuming this woman would have used it if one was available, so please correct me if I’m wrong.
These cars are likely going to turn into hijack machines if they’re programmed for “maximum safety” in situations where, realistically, breaking every traffic law, hitting a pedestrian or causing damage to the vehicle through dangerous terrain may be the only way out with a living passenger. The second it begins to percolate among criminals that these things are super easy to stop at the perfect location of your choosing like this, they are going to become a massive target.
Or they turn into a hearse if the passenger has a medical emergency and the car doesn’t redirect while the passenger is incapacitated. They might be coherent enough to press a button, but not to open their phone, navigate the app, call for help or redirect the car to a hospital…
But that of course requires labor so it will not happen until legally mandated after a minimum threshold of people die.
Nobody reads the article though…
Agreed, but to play devil’s advocate, the support wasn’t branded as such and customers could’ve not reported out of shame, which wouldn’t happen if they knew they could do that at the beginning before it became anything substantial.
Honestly a proper panic button would have an alarm go off and speed dial 911. But I’m sure people would abuse it.
She talked to an operator who asked if he should call the police and she said no. It’s in the article.
Not sure what a button would have changed…
It’s blocked for me unless i want to sign up. And I don’t for one article.
They have customer support that provides words of platitude, an ineffective police call with a 15minute response time, and no control over the situation. She got lucky this time, but my point remains standing.
They should have a support chopper that you can call in
Hello, customer support?..Yes, air strike on my location, thank you
I must be tired right now but I don’t see how a remote operator could have driven better in this situation.
You can’t get away from someone blocking your car in traffic without risk.of hitting them or other people or vehicles.
You probably meant they ought to drive away regardless of what they hit, if it helps the passenger escape a.dire.situation? But I have to wonder if a remote operator would agree to be put on the spot like that.
Yea I’m not too keen on giving authorization to hit pedestrians. If I feel threatened in my car, I am not allowed to run over the person so why should a driverless car gain that right? And if the panic button is going to call the police, how is that any different from the passenger using their phone to contact police? Seems like extra steps of middlemen and confusion when the passenger could just call once they feel the need.
I could defintely see a case for some extra safety features that help keep the doors locked and shut, maybe thicker windows too if needed to prevent robberies/assaults.
You are not allowed to run people over merely because you feel threatened.
You are allowed to use deadly force, in the USA when you reasonably believe that it is necessary to prevent someone from unlawfully killing, causing serious physical injury, or committing a short list of violent felonies. The harassment described in the article probably does not rise to that level, though an ambitious lawyer might try to describe intentionally causing the car to stop as carjacking or kidnapping.
Yeah somehow I don’t think tipping a fedora counts lol
It’s likely the harassers can be prosecuted for false imprisonment, a misdemeanor. It is illegal to use deadly force such as hitting people with cars to prevent/terminate a misdemeanor.
Is there any law in any state that would allow you to kill a 3rd party to escape being killed yourself? (If there were, I’d probably opt for not living in that state)
What do you mean by “allow you to kill a 3rd party”?
Like if rioters are breaking into your window and start trying to pull you out through it, then you floor it and kill someone else in the crowd who wasn’t actively breaking into your car?
This is something that’s going to vary from state to state, but ultimately it will be a case by case decision where a jury will decide if the use of deadly force was reasonable.
You will be judged based on other’s perception of the events, not based solely whether you yourself thought you were in danger or not.
So, someone trying to “drive slowly” through a group of protesters would probably be found at fault, while a car that was stuck trying to wait patiently suddenly having a Molotov cocktail thrown on it would be judged differently. Even then they will need to consider whether you could have just gotten out of your car and run.
https://www.reuters.com/article/world/fact-check-drivers-dont-have-the-right-to-plow-through-protesters-idUSKBN23B39F/
Sure, there are some states that let you mag dump through your front door if someone rings the doorbell
I vaguely remember reading in my criminal law textbook, years back, that murder is one of the few exceptions to the doctrine of necessity (this would have been in the context of US law), so I don’t think that it’s ever legally-permissible to intentionally kill some random person to save yourself. IIRC the rationale was that it prevents thing like terrorist groups from coercing someone to do actions for them by threatening someone else.
That being said, there are obviously points where people are forced to take actions where either one group of people is going to die or another; in ethics, the trolley problem is a well-known example. For a maybe-less-artificial problem, closing hatches in a ship where not everyone is out of a compartment to prevent the ship from going down, say. I don’t know how law applies in the situation of weighing lives; my assumption is that it doesn’t mandate inaction.
The “hitting pedestrians” is an extreme hypothetical, and not one you should particularly get hung up on. But it is one that still has to be considered. Passive security measures only go so far for the passenger.
Realistically, a car can get out of a vast majority of situations evasively without hitting hostile pedestrians, such as reversing rapidly and then turning around or driving in an opposite travel lane to bypass the blockage. Or hopping a curb and using a sidewalk if it is not occupied (or just blasting the shit out of the horn if it is occupied). These are all things that waymo’s auto mode cannot and will not do, because it doesn’t have the reasoning to understand when such measures are necessary.
If you legitimately believe your life is in danger, you have the right to escape or defend yourself, even if that means running someone over. This has happened in multiple countries with similar outcomes.
Think of it as a backup for the phone in the case where, say, there’s an adult and a kid in the car, the kid has no phone of their own, and the adult loses consciousness with their phone locked. Or the car is being actively jostled by a group of people (say it drove into the middle of an embryonic riot), causing the passenger to drop their phone, whereupon it slides under the seat. Or the phone just runs out of charge or doesn’t survive getting dropped into the passenger’s triple-extra-large fast-food coffee. It won’t be needed 99% of the time, but the other 1% might save someone’s life, and (presuming the car already has a cell modem it in) the cost of adding the feature should be minimal.
If you are in literal, actual mortal danger you are generally allowed to escape with the goal of escape. Especially relevant where waymo operates.
I can’t think of a NY cab driver that couldn’t have handled this situation.
This guy isn’t doing fedoras any favors either - I’m already a bit on the skeptical side when I see a fedora.
Fedoras haven’t done anyone any favors ever.
If a man jumps out in front of my car in traffic and points a pistol at me after I stop. I am going around or thru him and there is no other option. Anyone else trying to stop me even without visible weapons is going to get evasive maneuvers to protect myself because I am not dealing with that bullshit. That includes weaving far outside my travel lane or going over a sidewalk. That is self defense and a split second decision that any driver may have to make. Waymo prioritizes all outside obstacle avoidance which means it doesn’t even want to leave it’s set travel lane, which makes them trivial to stop like this with no recourse.
The point I am making is that self driving has a really hard time interpreting traffic edge cases or passenger emergencies like this. A remote operator could make the decision to drive over curbs and other lanes, if free, to save the passenger, and realistically should avoid hitting pedestrians too… but in the case of an armed attacker - well, yknow. Like force for like force.
Calling police would only be an auxiliary function to report the video evidence. They cannot be depended on to respond in time to actually make a difference.
Would a remote operator interpret things accurately in 10 seconds or less, or be a job anyone would even want? How does the liability chain of command work? Who knows. But the current system makes no decision at all, and that is unacceptable. And the medical point still stands too, a remote operator could immediately reroute the vehicle to a hospital and alert the medical staff. A panic button is absolutely needed.
I’m hitting them. I don’t know their intentions. But my intent would be to get away however I can.
What are you going on about? Have you ever ridden in one of these?
They do have these buttons…
https://support.google.com/waymo/answer/9172373?hl=en
It sounds like Waymo were already aware of the situation, in fact they called her in the vehicle as it was happening.
Not to say this isn’t a good suggestion, but they seem to have other systems in place that worked.
It worked, only because these men were only being creepy sexist pieces of shit and didn’t have worse intentions. Customer support according to this article has no control over the vehicle other than restarting the auto driving routines to make the car move again.
They do, she only used it after they were gone.
They have a button on the center-front thingy but it’s not labeled panic or anything.
Considering the length of your comment, you could have started by reading the article.
Well and the draw of these tiny driverless train like objects kinda goes out the window when you have to staff anything at all to monitor and control them.