Why Self-Driving Cars Must Be Programmed to Kill

Forums › Transport (cars, bikes and boats) › Why Self-Driving Cars Must Be Programmed to Kill

1 | ... | 8 | 9 | 10 | 11 | 12 | 13

gzt

17146 posts

Uber Geek

Lifetime subscriber

#1422807 7-Nov-2015 12:52

also it is not just control, but sensing beyond human capability.

Taking just a simple example, eye level and position vs roof based sensing giving extra distance.

Rikkitic

Awrrr
18663 posts

Uber Geek

Lifetime subscriber

#1422816 7-Nov-2015 13:03

Admittedly this was a long time ago, but when Apollo 11 landed on the moon, the computer panicked and screamed 'I'm going to die!', while the human saved the day.

Plesse igmore amd axxept applogies in adbance fir anu typos

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1422868 7-Nov-2015 15:17

Rikkitic: Admittedly this was a long time ago, but when Apollo 11 landed on the moon, the computer panicked and screamed 'I'm going to die!', while the human saved the day.

Not true, actually the computer completed its critical tasks despite being overloaded, due to good programming: https://en.wikipedia.org/wiki/Apollo_Guidance_Computer#PGNCS_trouble

Linuxluver

5828 posts

Uber Geek

Trusted

Subscriber

#1422959 7-Nov-2015 21:21

DizzyD: Found this article very interesting. What are your thoughts? Should self driving cars be programmed to kill?

Self-driving cars are already cruising the streets. But before they can become widespread, carmakers must solve an impossible ethical dilemma of algorithmic morality.

How should the car be programmed to act in the event of an unavoidable accident? Should it minimize the loss of life, even if it means sacrificing the occupants, or should it protect the occupants at all costs? Should it choose between these extremes at random? (See also “How to Help Self-Driving Cars Make Ethical Decisions.”)

The answers to these ethical questions are important because they could have a big impact on the way self-driving cars are accepted in society. Who would buy a car programmed to sacrifice the owner?

These problems cannot be ignored, say the team: “As we are about to endow millions of vehicles with autonomy, taking algorithmic morality seriously has never been more urgent.”

http://www.technologyreview.com/view/542626/why-self-driving-cars-must-be-programmed-to-kill/

The answer to almost all will be "pull over and stop".

Whatever happens....ba-bump, ba-bump....ooops.....it will pull over and stop.

_____________________________________________________________________

I've been on Geekzone over 16 years..... Time flies....

roobarb

653 posts

Ultimate Geek

Trusted

#1422968 7-Nov-2015 21:36

DizzyD: Found this article very interesting. What are your thoughts? Should self driving cars be programmed to kill?

Self-driving cars are already cruising the streets. But before they can become widespread, carmakers must solve an impossible ethical dilemma of algorithmic morality.

Let us summarise;

In order for X to do Y, Z must do the impossible.

Seeing as Z cannot do the impossible, X cannot do Y.

tdgeek

29750 posts

Uber Geek

Trusted

Lifetime subscriber

#1423031 8-Nov-2015 08:20

roobarb:
DizzyD: Found this article very interesting. What are your thoughts? Should self driving cars be programmed to kill?

Self-driving cars are already cruising the streets. But before they can become widespread, carmakers must solve an impossible ethical dilemma of algorithmic morality.

Let us summarise;

In order for X to do Y, Z must do the impossible.

Seeing as Z cannot do the impossible, X cannot do Y.

But before they can become widespread, carmakers must solve an impossible ethical dilemma of algorithmic morality.

This means that the cars will be driven by AI or near AI computers. When humans are confronted with an impending accident, do we decide this? " Uh oh, impending accident that I cannot avoid, so how much is a panel and paint repair, cost of those fancy LED lights, does the other car have full insurance, I am younger than that driver, so I might sacrifice his injuries for mine, I'll only cause minor injuries, so I might prefer to protect my 2 year old Aston Martin first."

Ok, a bit of exaggeration there, but when we are in that situation it all happens quick, and all we do is avoid and stop as much as we can. Thats all a CPU will do. We don't make ethical decisions when a car comes round an open road bend on our side of the road, or if another car runs a red light. We just apple basic rules of physics to avoid a crash if we can and stop as quick as we can.

roobarb

653 posts

Ultimate Geek

Trusted

#1423089 8-Nov-2015 11:23

tdgeek: We just apple basic rules of physics to avoid a crash if we can and stop as quick as we can.

We don't even do that, hence why people flip their cars in ditches on country roads trying to avoid rabbits. However the car still dutifully applies the basic rules of physics, hence the unintended end-state.

Quic Broadband

Move to New Zealand's best fibre broadband service (affiliate link). Free setup code: R587125ERQ6VE. Note that to use Quic Broadband you must be comfortable with configuring your own router.

tdgeek

29750 posts

Uber Geek

Trusted

Lifetime subscriber

#1423091 8-Nov-2015 11:33

roobarb:
tdgeek: We just apple basic rules of physics to avoid a crash if we can and stop as quick as we can.

We don't even do that, hence why people flip their cars in ditches on country roads trying to avoid rabbits. However the car still dutifully applies the basic rules of physics, hence the unintended end-state.

Very true, the genuine response which happens to be the wrong response. CPU control system wont make those human based errors

jarledb

Webhead
3257 posts

Uber Geek

Moderator

ID Verified

Trusted

Lifetime subscriber

#1424080 10-Nov-2015 02:17

Interesting about Tesla vs Google self driving cars on Quora

Jarle Dahl Bergersen | Referral Links: Want $50 off when you join Octopus Energy? Use this referral code
Are you happy with what you get from Geekzone? Please consider supporting us by making a donation or subscribing.

roobarb

653 posts

Ultimate Geek

Trusted

#1424161 10-Nov-2015 09:45

jarledb: Interesting about Tesla vs Google self driving cars on Quora

Very interesting, curiously no mention of the three laws of robots, ethics or calculating the impact of a crash.

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1424202 10-Nov-2015 10:48

I'm just going to leave this here

http://existentialcomics.com/comic/106

Rikkitic

Awrrr
18663 posts

Uber Geek

Lifetime subscriber

#1424273 10-Nov-2015 11:46

ubergeeknz: I'm just going to leave this here

http://existentialcomics.com/comic/106

You win. 42 plus ones.

Plesse igmore amd axxept applogies in adbance fir anu typos

DravidDavid

1907 posts

Uber Geek

#1424481 10-Nov-2015 15:58

I personally can't wait for self driving cars. Once they are mainstream and affordable, all Aucklanders will have them. Which means I'll get to drive around Auckland without having to worry about a crazy Aucklander that can't drive, because they won't be driving.

How cool would that be!

The semi-sarcasm aside, I feel once self driving cars are actually mainstream and affordable and people know how to behave around them, that accidents will become a rarity and ethical decision making won't need to be made by the computer in the long run. All ethical decisions aside, all the computer has to do is be better than a human behind the wheel, not necessarily perfect.

To decide if a computer would be better over and above a human, we have to put the computer in a situation where a human does not perform well.

A roadkill accident resulting in death (say a horse in the middle of the road on a foggy morning) is not necessarily the driver's fault. But it could almost certainly be prevented with a machine that can apply the brakes in 3 milliseconds over and above the normal reaction time of a human without considering the fact that a self driving car would almost certainly be able to detect the horse from a safe distance with radar, mitigating the flaws of the human eye.

Some might argue that most accidents were preventable, including the horse scenario when following the road rules. I've never been involved in an accident, while everyone else in my family have been in several. Having observed the level of care and attention my own family affords to driving suggests to me that perhaps a car that drives itself (regardless of ethics) would have a better way of dealing with an accident in almost all cases compared to an inattentive/distracted, incapacitated or irresponsible human.

With that said, what happens in the event that the software and hardware faults and the only occupant is in the passengers seat? How will insurance cope with having two drivers claiming not to be at fault? How would courts decide if the car or driver committed homicide or was guilty of a driving infraction if there is nobody at the controls?

--

TL;DR: Aucklanders suck at driving and human kind would be better if they all had self driving cars. Everyone should be worried about hardware and software faults causing accidents, rather than trying to find a way to build an ethics calculator.

1 | ... | 8 | 9 | 10 | 11 | 12 | 13