Hacker News Reader

Comment by pge | Hacker News Reader

This situation is not any different from a ticket issued by a speed camera. Since the camera can’t tell who is driving, the fine is issued to the owner of the vehicle (determined by the license plate). The owner is legally responsible for the fine. The only difference between a speed camera ticket and a ticket given by an officer in person is that the speed camera does not result in penalties for driving infraction (eg points on a license or license suspension) only fines. A ticket to a driverless taxi should work the same way; the citation goes to the owner of the vehicle.

View on HN

Is this really how it works in America? In the UK the notice of infraction is sent to the registered keeper of the car, obviously specifying where and then the offense occurred, and you then have the choice of telling the police who was driving the car or going to court. And, since it's your car and you're responsible for it, you'd better have a good reason for being unable or unwilling to say who was driving it at that time. There is no question whatsoever of a camera being unable to deliver points on the license.

Other than the fact that there are 50+ states here, and that the rules are different in each one, this is mostly correct. Ultimately corporations have a lot of privilege and deference compared to individuals. And traffic cameras are run by companies who often provide some sort of fee to the city.

This answer makes sense. But this is the law we're talking about, so I'm not sure about it.

this is incorrect (for us law). The operator of the vehicle is legally responsible for the ticket. An owner can show up in court and say “I don’t know who was driving my car at that time and place”. That’s why traffic cameras take a very clear photograph of the driver.

Assign one safety driver per vehicle, the remote safety driver is responsible (like a driving instructor and its driving student, except the AI is the student).

This way, the internal safety drivers will do everything they can to have the company act responsibly, and refuse the work if the system is not reasonably safe.

It can also solve other problems:

In Estonia we have delivery "robots". On the paper it is "AI", but in reality they are actually remotely controlled by students aka "Robot operators".

The law makes the other drivers guilty if you hit one of these "AI-robots", which really sucks, because the law was changed due to intense lobbying.

I'd rather have people designated as drivers to be responsible.

The point of Waymo is that human intervention is only used in exceptional circumstances. There isn't a one-to-one mapping on purpose, because it's not needed in a system generally controlled by AI.

Yes I agree, Waymo is real, and very nice tech, but if there is such high confidence, couldn't we find drivers who can remotely monitor the rides for safety ?

It could potentially solve the double-speech: "oh no no no, never willing to be the legal driver it's not a safe system", and at the same time: "oh but this system is very safe, we can put it in real traffic with real lives on the line".

Even if the system is safer than human drivers (I can’t speak to this), why would the lowest paid workers at the company voluntarily sign up for _any_ increased liability? If one of the programmers pushes out a bug that kills someone, why should an employee who had nothing to do with this be charged with a crime?

You raise a good point - if the system is so safe, the people who make such claims should accept some liability - but those are not the same people as the ones who sometimes intervene in the driving.

A separate job, remote safety drivers, who are responsible for monitoring the rides and acting (aka braking) in case of danger or mistakes.

Except from home / office.

You're asking for... what? For the secretary to be liable because a random car they have never seen or interacted with broke a law? How is that in any way shape or form reasonable? And why would anybody ever sign that contract except under duress or exceptional circumstances when they can get a job... literally anywhere else?

And even if that was a viable approach right now, the point of AI/self-driving technology is that the number of cars will vastly outnumber the number of employees once fully deployed and not in limited pilot programs.

The responsibility of this safety driver is to monitor the ride and be responsible in case of accident.

It's his main task, like any driver, just he has to pay attention to less things than a standard ride, since he has a very smart car (like when you sit in a Tesla with FSD).

Just he is not sitting in the car, instead in his office or home.

I basically covered this already in my previous comment, so I'm just going to restate it.

> why would anybody ever sign that contract except under duress or exceptional circumstances when they can get a job... literally anywhere else?

> And even if that was a viable approach right now, the point of AI/self-driving technology is that the number of cars will vastly outnumber the number of employees once fully deployed and not in limited pilot programs.

I understand what you mean, though drivers already take personal risks on the behalf of corporations.

If you are a Uber-driver, and enable the FSD on your Tesla, it's not Uber nor Tesla who will get in troubles if you are having an accident.

For now the situation with Waymo, can be the same as a driving instructor. The remote driver is the driving instructor, and Waymo AI is the student getting monitored and doing the driving.

The driver is responsible and, then they can potentially sue (or have agreement with Waymo) if it caused damages beyond their own responsibility (for example, if driver hits the brakes, and Waymo has overridden his manual commands).

Once this solution is safe, drivers will get more comfortable babysitting more than one car at once; essentially solving the scale issue that you mention.

Reply on news.ycombinator.com