Bikes May Have To Talk To Self-Driving Cars For Safety's Sake

[djg21]

Proponents of self-driving cars say they'll make the world safer, but autonomous vehicles need to predict what bicyclists are going to do. Now researchers say part of the answer is to have bikes feed information to cars.

http://www.npr.org/sections/alltechc...r-safetys-sake

This is interesting; the ethical dilemmas surrounding the so-called “internet of things” and self-driving cars in particular. What is a self-driving car programmed to do when faced with either a collision with an oncoming car or a collision with a cyclist?

[FlashUNC]

I'll bring back up what I mentioned last time we discussed autonomous cars round these parts: Self-driving cars are going to have to be programmed who to kill, basically.

[BobO]

Originally Posted by FlashUNC

I'll bring back up what I mentioned last time we discussed autonomous cars round these parts: Self-driving cars are going to have to be programmed who to kill, basically.

It'll still be orders of magnitude better than it is now with legions of dip****s invading the bike lanes while Snapchatting.

[David Tollefson]

The end-game is that roads will become "driverless car only" venues. Bikes will be banned -- relegated to specific facilities that go nowhere, or to trails (MTBs and gravel forest service roads).

[MattTuck]

Originally Posted by FlashUNC

I'll bring back up what I mentioned last time we discussed autonomous cars round these parts: Self-driving cars are going to have to be programmed who to kill, basically.

Yes, in an extreme situation, there needs to be some logic built in for this. But the rest of the time, the hope is that the car is less distracted than a human driver. If you're paying attention all the time, you're much more likely to avoid situations where the computer has to decide who to kill.

That said, I imagine it could be somewhat more comforting if a computer kills your family member, and the logs of that decision are preserved and open to analysis so you can understand the logic, ethical backing and logic behind it, as opposed to getting mowed down by a drunk.

[William]

Accident Report: Autonomous car collides with cyclist.

Investigation:
Data download from autonomous car:

Downloading...


Downloading...


Downloading...


Downloading complete.


Resulting response from autonomous car...



"I did not see him, the sun was in my camera eye."








William

[makoti]

Originally Posted by FlashUNC

I'll bring back up what I mentioned last time we discussed autonomous cars round these parts: Self-driving cars are going to have to be programmed who to kill, basically.

Will I be able to do that?

[FlashUNC]

Originally Posted by MattTuck

Yes, in an extreme situation, there needs to be some logic built in for this. But the rest of the time, the hope is that the car is less distracted than a human driver. If you're paying attention all the time, you're much more likely to avoid situations where the computer has to decide who to kill.

That said, I imagine it could be somewhat more comforting if a computer kills your family member, and the logs of that decision are preserved and open to analysis so you can understand the logic, ethical backing and logic behind it, as opposed to getting mowed down by a drunk.

I put some doubt into the ethical backing when its, say, Mercedes programming the car to save the Mercedes owner above all else. That's going to be a fun lawsuit to work through.

[Mark McM]

Originally Posted by FlashUNC

I'll bring back up what I mentioned last time we discussed autonomous cars round these parts: Self-driving cars are going to have to be programmed who to kill, basically.

Sure - although the actual mechanism will probably be that cyclists are specifically targeted, but simply left off the list of "who not to kill".

But maybe there's hope. Computer programs are generally good at being predictable. Once the programming is understood, maybe cyclists can "hack the system" - maybe one scheme would be finding some way of tricking the car into thinking that the cyclist is a celebrity, or a nun, or somebody that the car specifically is programmed not to kill.

[Mark McM]

Originally Posted by FlashUNC

I put some doubt into the ethical backing when its, say, Mercedes programming the car to save the Mercedes owner above all else. That's going to be a fun lawsuit to work through.

The driving computer will be able to do an instantaneous calculation on which death will have the least financial liability.

[54ny77]

that poses an interesting legislative and legal conundrum.

....and a new curriculum for philosophy professors to keep that tenure hope alive!

Originally Posted by FlashUNC

I'll bring back up what I mentioned last time we discussed autonomous cars round these parts: Self-driving cars are going to have to be programmed who to kill, basically.

[Mark McM]

Originally Posted by William

Accident Report: Autonomous car collides with cyclist.

Investigation:
Data download from autonomous car:

Downloading...


Downloading...


Downloading...


Downloading complete.


Resulting response from autonomous car...



"I did not see him, the sun was in my camera eye."








William

From the movie 2001: A Space Oddysey:

HAL 9000: Well, I don't think there is any question about it. It can only be attributable to human error. This sort of thing has cropped up before, and it has always been due to human error.

[FlashUNC]

Originally Posted by Mark McM

From the movie 2001: A Space Oddysey:

HAL 9000: Well, I don't think there is any question about it. It can only be attributable to human error. This sort of thing has cropped up before, and it has always been due to human error.

HAL actually ain't that far off. A computer that was infallible, until given contradictory instructions by his makers to the classified nature of the Jupiter mission.

[fiamme red]

Originally Posted by Mark McM

From the movie 2001: A Space Oddysey:

HAL 9000: Well, I don't think there is any question about it. It can only be attributable to human error. This sort of thing has cropped up before, and it has always been due to human error.

HAL 9000: Let me put it this way... The 9000 series is the most reliable computer ever made. No 9000 computer has ever made a mistake or distorted information. We are all, by any practical definition of the words, foolproof and incapable of error.

[marciero]

Originally Posted by FlashUNC

I put some doubt into the ethical backing when its, say, Mercedes programming the car to save the Mercedes owner above all else. That's going to be a fun lawsuit to work through.

I understand the rationale for Mercedes decision to protect the driver is that consumers would not buy cars programmed to for example, cause the least harm, even at the expense of the driver. The argument is that more people would buy them, resulting in more self-driving cars on the road. More on the road results in safer environment for all, which I agree with 100%. Thus programming cars to protect the driver would be result in safer roads.

Originally Posted by 54ny77

that poses an interesting legislative and legal conundrum.

....and a new curriculum for philosophy professors to keep that tenure hope alive!

Indeed. There are some interesting dilemmas that can arise when the car is programmed to do the least harm. Some of these are not new at all. Would that be "least harm" to the the individual or defined in some other way? What should the car do given the choice between maiming a 20-year old and killing an 85-year old? Between hitting a cyclist without a helmet and one with a helmet? The helmeted cyclist is protected so would be likely to injured less seriously. But should the cyclist wearing the helmet be penalized for making the responsible choice to wear a helmet? (Just for the sake of argument- No intent to start on helmets!) If that is the case, would it perhaps be safer not to wear a helmet, knowing that cars would be less likely to hit you? But if no one wore a helmet no one is safer.

(BTW the last example is an example of a "prisoners dilemma" on the part of the cyclist)