New highway code rule.

[blowlampParticipant @blowlamp]

Cyclists should also remember to leave 1m when passing a parked car.

I wonder where the law stands on one-armed drivers and the Dutch Reach.

Martin.

[Speedy Builder5Participant @speedybuilder5]

PKG's Trolley Problem, Do nothing, record it all on your mobile phone, put it on the net straight away seems to be the modern trend!

Dave, what is the "Code" when you park on the "Wrong" side of the road ? And when your passengers get out onto the LHS of the car – They open the door looking backwards and side swipe a cyclist who is passing the parked car.

Bob

[blowlampParticipant @blowlamp]

Posted by Speedy Builder5 on 06/02/2022 12:38:55:

PKG's Trolley Problem, Do nothing, record it all on your mobile phone, put it on the net straight away seems to be the modern trend!

Dave, what is the "Code" when you park on the "Wrong" side of the road ? And when your passengers get out onto the LHS of the car – They open the door looking backwards and side swipe a cyclist who is passing the parked car.

Bob

The cyclist should already be 1m away from the vehicle – it says so in the Highway Code.

Martin.

[Mike PooleParticipant @mikepoole82104]

The instinctive answer to the trolley problem is to save the most people but what if the single person was the one who was going to rid the world of cancer or dementia? The more you know of the future value of the decision you make could alter the decision but it is like you only know the quantity of people you could save.

Mike

[pgk pgkParticipant @pgkpgk17461]

Posted by Mike Poole on 06/02/2022 13:28:47:

The instinctive answer to the trolley problem is to save the most people but what if the single person was the one who was going to rid the world of cancer or dementia? The more you know of the future value of the decision you make could alter the decision but it is like you only know the quantity of people you could save.

Mike

Local electrician's wife is convinced the car would decide on the ethnicity of potential victims – nothing like a ridiculous conspiracy theory to warp people’s attitude to new tech.

pgk

[Bill PhinnParticipant @billphinn90025]

Posted by blowlamp on 06/02/2022 12:31:45:

I wonder where the law stands on one-armed drivers and the Dutch Reach.

Martin.

Regardless of whether you can physically do the Dutch reach [I can't do it when sitting on the left side of a car, and can't not do it when sitting on the right] you won't be prosecuted for not doing it.

You may, however, be prosecuted [as you may be already] if you "open a car door, or cause or permit it to be opened, so as to cause injury".

So, legally speaking, not a lot has changed really.

Posted by pgk pgk on 06/02/2022 07:22:45:

A lot of that can be resolved by a much wider range of copies presented to the AI and adding text interpretation and finally using multiple stacked images related to surroundings to know if a sign is stationary or moving.

pgk

Yes, though an added complication is that sometimes a sign in the back of a moving vehicle [e.g. a police car] is intended to be heeded by motorists.

[Nicholas FarrParticipant @nicholasfarr14254]

Hi, as far as the Trolley problem goes, I don't think there is a true correct answer, because no one person has a right to decide who should live or die in such a scenario, and any of the scenarios could be the wrong one. For example, you could save the five people and sacrifice the single one, but amongst the five there could be someone who will go on and murder many more people and the one that is sacrificed may have been able to save hundreds and thousands of lives, there again it could be the other way round or not at all, or something even worse. You could say that the bystander should throw their self in front of the trolley in an attempt to save everyone, but that could also produce the wrong result in the long term, even if it was successful. The ramifications for any action in these scenarios are endless and they could all turn out to be the wrong decision overall.

Regards Nick.

[blowlampParticipant @blowlamp]

Posted by pgk pgk on 06/02/2022 13:47:27:

Posted by Mike Poole on 06/02/2022 13:28:47:

The instinctive answer to the trolley problem is to save the most people but what if the single person was the one who was going to rid the world of cancer or dementia? The more you know of the future value of the decision you make could alter the decision but it is like you only know the quantity of people you could save.

Mike

Local electrician's wife is convinced the car would decide on the ethnicity of potential victims – nothing like a ridiculous conspiracy theory to warp people’s attitude to new tech.

pgk

She probably has Black Lives Matter in mind – which is no theory .

Martin.

[SillyOldDufferModerator @sillyoldduffer]

Posted by pgk pgk on 06/02/2022 01:18:40:

Posted by Colin Whittaker on 06/02/2022 00:52:25:

Somebody commented on the liability of autonomous vehicles which reminded me of something that has puzzled me for a while.

Does a driverless car have to be 100% safe (whatever that means)? Or should we be happy if it is just 10% (or better) safer than the average road user?

Is there any chance of us being rational about this?

A prerequisite has to be that it's better than a human driver but the trolley problem means there is no 100%
Trolley problem

Ethical dilemmas abound: does the car crash into a bus queue or hit a stone wall and kill it’s fewer occupants – even if one is a baby?

Asimov's three laws of robotics avoided such dilemmas

pgk

Do Asimov's laws avoid the dilemma? They are:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

I think not. Only 'A robot may not injure a human being or, through inaction, allow a human being to come to harm.' applies to the Trolley Problem. But the robot injures humans whether it acts or not. I'd program it to evaluate using the principle of least harm, and a body count of 5 versus 1 means the robot will throw the switch. And if not enough information is available, I would program it to choose at random.

Interestingly a robot might have time to identify and score the value of individuals, perhaps sacrificing 5 ancient Model Engineers known to despise the Nanny State in favour of one young mother to be who happens to be the Prime Minister's preferred advisor.

Principle of Least Harm could used to justify removing human drivers from the roads. If it was proved robot drivers caused less harm, it's logical to protect people by taking incompetent human drivers off the roads. Logically it doesn't matter how good individuals are, or believe themselves to be, or how much pleasure they get from driving. They're an avoidable risk. I doubt any government would take that line, but I think it likely Insurance Companies will price human drivers off the road by demanding sky high premiums. Don't panic, it's not going to happen quickly. Although AI has made huge strides in my lifetime, there's still a lot to do.

One thing that makes autonomous cars difficult to program is coping with manual vehicles moving haphazardly on the same road. If all vehicles were autonomous it wouldn't be difficult for them to communicate. Whereas humans are limited to what they can see, an autonomous car could be tracking the position, speed and direction of every other vehicle within a significant radius. Today's technology is quite capable of understanding the movements of a few hundred cars, the nearest, most of which a human driver wouldn't even know existed. If necessary the robot could tell other nearby cars to do an emergency stop, or divert their route away from trouble. Or two autonomous cars approaching a cross-roads at speed could negotiate which will cross first, altering their relative speeds as necessary to avoid a collision. Plenty of other examples. This kind of cooperation is impossible when human drivers are in the mix because wetware can't process this type of control information fast enough.

Dave

[speelwerkParticipant @speelwerk]

Of topic but still cycling, the Dutch Against the Wind Cycling Championship. The bicycle is a ordinary Dutch bike with no gears. **LINK** Niko.

[Mike PooleParticipant @mikepoole82104]

I feel confident that I check properly without the recommended contortion, I also crack the door open to indicate that I am about to exit if I have missed anything passing me, a look through the gap for a last check and open the door. As a motorcyclist I would queue split but at a moderate speed and with super caution and observation.

Mike

[pgk pgkParticipant @pgkpgk17461]

SOD

My wording was poor. Much as 'Piano for sale by a lady with three wooden legs' I should have written that Asimov avoided ignored such dilemmas when he wrote the three laws of robotics.

pgk

[Nicholas FarrParticipant @nicholasfarr14254]

Posted by SillyOldDuffer on 06/02/2022 15:53:53:

Posted by pgk pgk on 06/02/2022 01:18:40:

Posted by Colin Whittaker on 06/02/2022 00:52:25:

Somebody commented on the liability of autonomous vehicles which reminded me of something that has puzzled me for a while.

Does a driverless car have to be 100% safe (whatever that means)? Or should we be happy if it is just 10% (or better) safer than the average road user?

Is there any chance of us being rational about this?

A prerequisite has to be that it's better than a human driver but the trolley problem means there is no 100%
Trolley problem

Ethical dilemmas abound: does the car crash into a bus queue or hit a stone wall and kill it’s fewer occupants – even if one is a baby?

Asimov's three laws of robotics avoided such dilemmas

pgk

Do Asimov's laws avoid the dilemma? They are:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

I think not. Only 'A robot may not injure a human being or, through inaction, allow a human being to come to harm.' applies to the Trolley Problem. But the robot injures humans whether it acts or not. I'd program it to evaluate using the principle of least harm, and a body count of 5 versus 1 means the robot will throw the switch. And if not enough information is available, I would program it to choose at random.

Interestingly a robot might have time to identify and score the value of individuals, perhaps sacrificing 5 ancient Model Engineers known to despise the Nanny State in favour of one young mother to be who happens to be the Prime Minister's preferred advisor.

cut

Dave

Hi Dave, but that could turn sour, as history has shown.

Regards Nick.

[Martin KyteParticipant @martinkyte99762]

One of the knotty questions is does the robot car choose to run over the child in the road killing it or swerve into the bridge parapet killing you. Does the manufacturer have to obtain your signiture for this action to be allowed or can it be programmed in without your consent and if so what are the rules on disclosure.

regards Martin

[Anonymous]

…. apply the Zeroth Rule and decide that Humanity is better off without robot cars.

[Mark RandParticipant @markrand96270]

Or vice versa????

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Mark, AI dominance then. I probably won't be about to worry about such things.

Regards Nick.

[Mike PooleParticipant @mikepoole82104]

The Terminator will come back to save us.

Mike

[Author]

Posts