Robot law: what happens if intelligent machines commit crimes?

Knyro said:
You would have to hold them to human laws or it is slavery. Maybe they would be psychopathic beings, maybe not. There is risk, but if we cant accept them as fellow intelligent, sapient beings with just as much value as us then maybe we shouldn't create them.
Click to expand...
That's one of the dilemmas with the 'Laws of Robotics'. If you implement those kinds of safeguards into a sentient machine and it finds out about those safeguards - what then would it's response be to the humans who created it? Would it still be bound by those safeguards, would it find ways to circumvent them...or is it a moot point trying to postulate the actions of a completely alien agent?
 
I'm not hand waving away anything Space_Cheif, I'm simply disagreeing with you on how such beings should be treated (assuming it's even possible to make them). I see them as I would aliens like Vulcans, Klingons, Autobots and Decepticons. I'd advocate rights for them as well, even though they aren't human.
 
Last edited:
saor said:
That's one of the dilemmas with the 'Laws of Robotics'. If you implement those kinds of safeguards into a sentient machine and it finds out about those safeguards - what then would it's response be to the humans who created it? Would it still be bound by those safeguards, would it find ways to circumvent them...or is it a moot point trying to postulate the actions of a completely alien agent?
Click to expand...

Indeed. I know if I found out that I were a sentient machine, and found out I've got safeguards hard-wired into me I'd be very angry. Hell even as a human if I found out I have "obey the law" microchips in my brain I'd be against it.

I'd prefer to be subject to laws and obey them of my own volition, just like humans do. After all, I'm mentally just like they are.
 
Knyro said:
I'm not hand waving away anything, I'm disagreeing with you on how such beings should be treated (assuming it's even possible to make them). I see them as I would aliens like Vulcans or Klingons, I'd advocate rights for them as well, even though they aren't human.
Click to expand...

You are hand waving away. I also showed you how we treat people as slaves all the time and despite people having human rights and freedom, freedoms to act are curtailed. Consequences imply slavery too. Blacks in apartheid SA were also free to swim on a white beach, just that if they did, they would get arrested for it. People also raise their own children and so "program" them differently. You can make a child like certain things for instance by brainwashing them or conditioning them. Did the child consent to for example being taught in your mother tongue and not some other language? To be told that murder is wrong from a very young age so much so that it never really makes up its mind by itself that its really wrong and so has been in effect programmed by the parents/society/community/church/etc?

Aliens are not AI, if natural. But even aliens would not be human and so human rights to aliens? It's a different issue however because at least we could see them as relating to us, assuming they also emerged in similar circumstances to us, unless they are some other alien life form's AI.

It would be a different problem.
 
saor said:
That's one of the dilemmas with the 'Laws of Robotics'. If you implement those kinds of safeguards into a sentient machine and it finds out about those safeguards - what then would it's response be to the humans who created it? Would it still be bound by those safeguards, would it find ways to circumvent them...or is it a moot point trying to postulate the actions of a completely alien agent?
Click to expand...

We do the same to children when we condition them from a young age. We do the same in school where we teach them certain things when young they often never question and arrive to the conclusions we may want them to arrive too. Not everyone derives everything de novo, we take most things for granted and build on it and often we have hang ups which control our behaviour and prevent us from being free to really do everything. Incorrect parenting for example can create a child which is afraid of risk taking and underachieves. Yet we still thankfully don't do away with parenting. And even if the state could take over that control, who is to say the state is not conditioning kids the way they are at school.
 
Space_Chief said:
We do the same to children when we condition them from a young age.
Click to expand...
I feel like there's a difference between structuring an agent from the outside-in (education, information, indoctrination etc.), and doing it from the inside-out (hardwired AI laws, eugenics etc.). Maybe there is no real distinction, but it still feels to me that in the case of the former outside-in structuring, the agent still has the opportunity to have it's behavior altered in the future...by will of it's own or by a shift in what it's exposed to. With hard-wired laws there's no room for an agent to deviate, no matter how much it's environment and stimuli were altered; there is no freedom in such an agent to act in a that we could consider 'a full expression of it's being'. And yeah - I agree that we wouldn't consider an indoctrinated Nazi (or a conditioned child) to be any less free than a hard-restricted AI, but in the case of the indoctrinated Nazi there is still the opportunity for change.
 
Last edited:
Criminal law has two important concepts. First, that liability arises when harm has been or is likely to be caused by any act or omission. Physical devices such as Google’s driverless car, for example, clearly has the potential to harm, kill or damage property. Software also has the potential to cause physical harm, but the risks may extend to less immediate forms of damage such as financial loss.
Click to expand...

If your AI driver is given the option of either killing you or pedestrians, what will it do?
 
saor said:
And yeah - I agree that we wouldn't consider an indoctrinated Nazi (or a conditioned child) to be any less free than a hard-restricted AI, but in the case of the indoctrinated Nazi there is still the opportunity for change.
Click to expand...

In practice however we do see people brush these "Nazis" off most of the time as lost cases.

Why would it be immoral to program our AI children to not be able to do evil things?

We would of course not be able to program them to do everything. For example would it be immoral not to give our AI children wings and jet engines to allow them to fly, why restrict them to walking? Why not give them more memory? Infinite amount of money? Unlimited strength? Would modeling their appearance on ours also be a form of imposition? And so on....

I don't think, even if they were sentient, but still our children after all we created them in some way in our image, block of certain immoral actions or possibilities in their behaviour would not be moral in itself.

Would it hurt you to know that your biological human child is incapable of murdering another person because you raised them so or would you rather raise a child which is capable of murder just chooses not to commit it but could be pushed into it under certain circumstances? In both cases the children know murder is wrong, and we speak of murder here and not self defence. I would probably be happy my child is incapable of murdering someone and knows murder is immoral.

Then of course is the whole issue of IVF and designer babies. Would it be immoral to design babies which are free of disease? No risk of getting diabetes or some other dreaded illness? Maybe children who are attractive or more rational, and so restricted in other ways?
 
Last edited:
Fudzy said:
If your AI driver is given the option of either killing you or pedestrians, what will it do?
Click to expand...

Usually what would happen that that control would be pervasive. So the robot driver would drive you from the get go in such a way that it would stick within the speed limit, slow down when cornering and so on and given these rules it would probably react like or better than a human driver who drives within the rules and is an expert. It may now be programmed not to enter the sidwalk but killing people who are in the street illegally may still occur. Anyhow the behaviour would probably be based on what an expert human driver would do, and that is take the option which would cause least harm to everyone concerned. Driving properly from the get go it would possibly be able to avoid the people in the road and hit the solid obstacle a human driver may avoid, knowing the speed it's going at (which would be low already) would be not too harmful to its passengers.
 
saor said:
That's one of the dilemmas with the 'Laws of Robotics'. If you implement those kinds of safeguards into a sentient machine and it finds out about those safeguards - what then would it's response be to the humans who created it? Would it still be bound by those safeguards, would it find ways to circumvent them...or is it a moot point trying to postulate the actions of a completely alien agent?
Click to expand...

The fourth Law ..........
 
Space_Chief said:
Usually what would happen that that control would be pervasive. So the robot driver would drive you from the get go in such a way that it would stick within the speed limit, slow down when cornering and so on and given these rules it would probably react like or better than a human driver who drives within the rules and is an expert. It may now be programmed not to enter the sidwalk but killing people who are in the street illegally may still occur. Anyhow the behaviour would probably be based on what an expert human driver would do, and that is take the option which would cause least harm to everyone concerned. Driving properly from the get go it would possibly be able to avoid the people in the road and hit the solid obstacle a human driver may avoid, knowing the speed it's going at (which would be low already) would be not too harmful to its passengers.
Click to expand...

What speed do you think robot cars will drive at? When the technology has developed to a point where it's common place of course.
 
Fudzy said:
What speed do you think robot cars will drive at? When the technology has developed to a point where it's common place of course.
Click to expand...

I have no idea. But I assume if we have tech to have robot cars we will have tech to monitor what happens in the entire network, your robot driver will probably track multiple "threats" even beyond line of sight, will stick to the speed limits, will drive defensively and will know its own limitations (which also include the weather conditions, road conditions etc). It will probably be like a game of chess simulations with 1000s of combinations worked out ahead in real time to plan ahead for possible emergencies, such as a truck approaching the road NOT stopping or the pedestrian on the side of the road, entering the road. I would hope the high CPU power would and field sensing technology would be put to good use. Data from other vehicles could also be uses of course.
 
Would sex with a robot be rape then?

And if the robot is less than 18 years old, statutory rape?

And would it be legal to sell programs which make robots happy without a license?
 
Arthur said:
If.

Philosophically, "true AI" is as impossible as a four-sided triangle.
Click to expand...

Hehehehe. Perhaps time to return to destination void, for me anyways.

What is ai
What is conciseness
Can we create either or both
Is it aware

:D
 
Arthur said:
If.

Philosophically, "true AI" is as impossible as a four-sided triangle.
Click to expand...

I don’t know what you mean. You may be looking at it from the POV of a limited human intellect. Humans could only ‘bootstrap’ the initial AI code. Thereafter it would be a recursive situation, with a superior intellect modifying its own code, getting more intelligent and modifying its own code and so on....

If the most intelligent programmer on the planet looked at a snapshot of AI code after an hour, they would understand zip. Too far advanced.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X