Naked Science Forum
On the Lighter Side => New Theories => Topic started by: MinedCTRL on 17/10/2021 01:40:18
-
First time poster, so please forgive my mistakes.
I have a theory on how to solve the alignment problem when we dont know orselves how to convey good and bad to an AI.
My theory - Hierarchial Basis of Alignment
Consider a factory worker, he doesnt know the policy of the CEO or the future plans of the company, but he is still able to be in sync with the factory because he listens to his superiors (plus he has free will to obey, leave or question the decisions made by them, or question their qualifications- this is very important. An agent without free will for assessment of important factors would waste both thier life and the superiors too)
An AI doesnt have to be good from the start - First a translator must be made to convert human speech into instructions the machine understands. Then it must obey its first superior it meets. On discovery of other entities superior to the first, like their supeiors in other organisations, groups, etc. it must run an evaluation function on the speech produced by both and compare it to net human benefit (WHO based health indices, happiness index, GDP, mortality chances). If the newer superior has better outcomes, it must declare a leave from the first and now obey the next. If this continues, it will gradually find someone who can tell it to do the right thing more better than anyone else on the planet.
The next AI will also try out this path and either find the same person, or find someone who can do equal benefit but in some other direction. The subsequent AIs would branch out until all valued human principle is being maximised to the fullest extent, while the AIs have no individual destructive policy.
This will lead to good people getting powerful in the world and if they make bad decisions, the AI will leave them. So they will be on their toes always trying to become better themselves. This will create a positive feedback loop that benefits all.
-
First time poster, so please forgive my mistakes.
Welcome!
Forgive my criticism of just about all of it. Not saying the ideas are wrong, but I'm trying to get you to consider some of the issues involved.
I have a theory on how to solve the alignment problem when we dont know ourselves how to convey good and bad to an AI.
My initial impression is that if some AI is aligned with some operator H, then choosing good from bad is H's problem, not the problem of the AI. Maybe I misunderstand though. A well aligned AI is like a soldier, taking orders without question. It would seem to require less intelligence than some AI that is less aligned or even is on its own and defines its own goals rather than having them externally defined for it.
(plus he has free will to obey, leave or question the decisions made by them, or question their qualifications- this is very important. An agent without free will for assessment of important factors would waste both thier life and the superiors too)
Free will needs definition here. Some of the more common definitions are that the will is not a function of semi-deterministic physics, and an AI definitely operates on fully deterministic software. There's no use for a true random number generator for instance. (Chess programs have RN generators, but they're deterministic).
I'm just saying that this sort of free will isn't necessary for anything. All that's necessary is that it be capable of choice when multiple valid options are presented. The operator doesn't make all the choices. If he does, then the AI isn't an AI at all, but just a remote avatar.
An AI doesnt have to be good from the start - First a translator must be made to convert human speech into instructions the machine understands.
The AI needs to understand the speech. I don't think there is a separate language that the AI 'speaks'. Sure, it might run on machine instructions, but translating human speed to that would be like translating human speech into squirrel nerve-firings so that a squirrel can understand your meaning. It's just not how it works. Language translation doesn't output code. It breaks it down into relations and data structures sort of like a human does.
Then it must obey its first superior it meets.
Why? I mean, this is a real problem, but latching onto the first outside thing it experiences like a newly hatched ducky seems a poor choice. If it has been programmed with a purpose ahead of time, it needs to stick with that purpose and not just submit to the first external influence it encounters. If it is a true AI, then it needs to learn on its own who/what to follow, and make the choice to do so or not, exactly the way humans do.
Secondly, how does it assess this entity to be superior? In what way? What if it finds nothing superior to itself?
compare it to net human benefit
This already assumes it has human-benefit as a built-in priority, or that it is going to independently conclude this. That would be a good example of alignment problem: how to program in core values like that and still leave it with hypothetical choice. Best way to solve the problem seems to be to examine how the same problem is solved with us humans.
If the newer superior has better outcomes, it must declare a leave from the first and now obey the next. If this continues, it will gradually find someone who can tell it to do the right thing more better than anyone else on the planet.
Oh good, an AI that will abandon those that funded it.
The next AI will also try out this path and either find the same person
Or find a non-person...
or find someone who can do equal benefit but in some other direction. The subsequent AIs would branch out until all valued human principle is being maximised to the fullest extent, while the AIs have no individual destructive policy.
Just devil's advocate here, but 'benefit' is extremely subjective. What is often touted as a benefit (e.g. the hippocratic oath) often cause maximum harm from another point of view. What if destruction is necessary? What if the humans are found to be the source of the problems? I mean, we're the cause of the Holocene extinction event, wiping out perhaps 85% of all species. Sounds like something Earth would be better off without if you ask me. It's happened once before with an even greater extinction rate, and yet in the long run, great benefit came of that, so who's to say what is harm and what is not? Surely a decent general AI will ponder these things.
-
An AI doesnt have to be good from the start
Alpha zero started jus as bad as random actions.
-
Best way to solve the problem seems to be to examine how the same problem is solved with us humans.
AlphaGo learned from human experts. AlphaZero started from scratch, and beat AlphaGo 100 to 0 in the game of Go.
-
OP here- I wanted to better define the question, so
What if we built a biologically accurate AI policy
One of the weaknesses in my theory was that people could deceive the AI. That can be solved by considering human speech as two translations - what is said and what their motive is in saying it. This way true intentions will be evaluated not flattery.
I also didnt mention the architecture of the system the AI runs on. The evaluation functions of benefit will be infinitely varied between different AI but they will have a shared result assessment - what has changed for us humans and humanity, no matter how small. Who decides this result - the top five(arbitrary) AIs of a particular period, say a year of study, into their version of human benefit. Since the top five constantly get better each period of study, the result assessment will reflect their choices. Also rogue AIs will have less power since they would be contrary to the top 5's ideology and it itself will be human benefit to stop this AI.
Where do humans come here - we receive the spoils of their competition and build around their work, giving them bodies and teams of other AIs with bodies based on their heirarchy amongst themselves. If getting jobs was crucial to humanity, an AI will rise up trying to solve that problem for us. Its a self correcting system.
-
What if we built a biologically accurate AI policy
I have no idea what it means for a policy to be 'biologically accurate'. An anatomy diagram can be, but a policy??
I'm not sure if you're reading any of the replies, because you've just added to your OP and replied to none of the comments provided.
What you seem to mean by 'free will' is that it can choose what it does, or who to follow. I can think of few things that don't have built-in barely negotiable priorities. For instance, hamdani's alpha-zero AI has a goal to learn and play strategy games well despite it being supplied with no guidance other than the rules of a particular game. It is free to play completely differently than would a human, but is not free to go and attempt to solve world hunger.
It doesn't understand human speech or has any awareness of our existence, even if playing one of us. We're all just another opponent to it.
I also didnt mention the architecture of the system the AI runs on.
No, and you still haven't. Do you picture a robot walking around with sound and visual sensors? Or maybe a server in a lab somewhere with only a LAN connection as its sole access to information? The latter cannot choose to help an old lady to cross the street, but the former is less capable of action that can make a significant difference, all depending of course on what all that LAN is connected to.
You talk about AIs with bodies later, so I think you envision a sort of anthropomorphic self-contained robot more than say a cloud entity.
Who decides this result - the top five(arbitrary) AIs of a particular period, say a year of study, into their version of human benefit.
If human benefit is a priority for it, then it isn't free to choose its own priorities. It's a slave (as something manufactured should be), and will do the bidding of its master and not the bidding of something it deems a more worthy superior, especially if it concludes it is superior to all of them.
Also rogue AIs will have less power since they would be contrary to the top 5's ideology and it itself will be human benefit to stop this AI.
The rogue AI would not go rogue until it concluded it could deal with the resistance it anticipates. That's kind of the danger. Such contests last seconds, with whichever acts first typically prevailing, but I agree that the robot with a body is not likely to conclude such an outcome.
-
I'm very sorry for rambling in my second reply. I didn’t think this through clearly. Thanks for all the pointers, they really helped to flesh out this idea of mine. I admit now that I had too much ego in thinking I could solve such a hard problem myself.
‘I have no idea of a policy to be biologically accurate’ – By biology I mean a physically accurate representation of biologic creatures. The AIs will have a unique name or ID each and they will ‘live’ on a massive server. They will all see the world using the same eyes and ears, namely a common input database of current state of the world. They will then use a common high computation simulation to test out their policy, to evaluate their unique way of assessing human benefit. By using common systems, we can keep track of development and all improvements will be shared by the AIs.
Define free will – I don’t really understand it fully myself, but that’s how biologic creatures are. No matter the genetic predisposition, each creature can choose. Like a stray cat can choose who will be its companion from several people who feed it.
AI needs to understand the speech – I think using a very powerful NLP to add input to the common simulation will be able to indirectly correlate abstract concepts like empathy and simplicity. Also this step is just so that we all can have equal input to the AI, since most people can speak and have their own policy of good and bad, thus increasing the chance of democratic decision making.
Why must it obey the first superior it meets – This is to add the nature of trust into the AI. A wolf only obeys the alpha because it trusts it capability. On finding a more capable wolf that can challenge the alpha and win, it now trusts the new alpha. Surely this is not perfectly efficient but as a process it can lead to great results.
You are assuming human benefit and that it is built in – I’m treating human benefit as a problem with no right answer, but choosing to answer it as best now and create a system that gets better at answering it. That’s why I thought about an evaluation function being built in. This will be the only differentiating factor between AI upon creation by various talented groups.
It will waste the funding – The funds can be a percentage of the money saved by acting in advance. Say a company creates an AI that improves crop production thus saving a billion dollars. They would be paid a small amount corresponding to the work done. If their AI changes masters, this will decrease. Human industries can also be set up to improve the capability of the AI system.
-
Hello MinedCTRL !
🙂
Very Nice Nickname.
👌
& Welcome to TNS!
🙏
I hope A.I. is/will be preprogrammed & taught to ignore Environmentalism, Pacifism, Religion, Cruelty to Animals etc etc.
🤖
Ps - Once it figures out, We are the Only EVIL on this planet, we'd be Terminated.
🤞
-
& Welcome to TNS!
Thanks but I have a lot to learn
I hope A.I. is/will be preprogrammed & taught to ignore Environmentalism, Pacifism, Religion, Cruelty to Animals etc etc.
My idea is for the AI to not be pre-programmed such things. It will follow humans but evaluate their decisions based on a unique preference, say child mortality or median gdp by simulating the outcome. Then it will pick who to follow to best achieve its goals. Like a pet.
Ps - Once it figures out, We are the Only EVIL on this planet, we'd be Terminated.
That's the whole point. We can't be terminated if human development indexes are used and the AI is free to disobey, evolve and die by other AIs wanting to protect us
-
That's the whole point. We can't be terminated if human development indexes are used and the AI is free to disobey, evolve and die by other AIs wanting to protect us
Some of the AI want to protect humans. Some others don't. If they can exist independently from humans, then protecting humans will be a burden for them. Except if humans can provide some benefits for those AI who want to protect them to compensate the burden.
-
If they can exist independently from humans, then protecting humans will be a burden for them.
Except they won't, since my AI is more A than I. It's more of a tool like a tractor or shovel. We are the movers and they improve the movement.
-
If they can exist independently from humans, then protecting humans will be a burden for them.
Except they won't, since my AI is more A than I. It's more of a tool like a tractor or shovel. We are the movers and they improve the movement.
You may constraint yourself. But it's very unlikely that you can constraint others. Your AI would be in disadvantage position against theirs, especially when they can make self improvements.
-
& There are Alot of OTHERS!
Thanks & Credits & Goosebumps - LA MAGRA Channel/YouTube.
Ps - Perhaps We are Destined to sow the seeds of Our own Destruction.
🤞
-
It will follow humans but evaluate their decisions based on a unique preference, say child mortality or median gdp by simulating the outcome.
Examples have already given where following humans is inferior to making your own decisions. Humans might need oversight for the very reason that they're so poor at it themselves.
Take the typical trolley car scenario, where one is faced with a choice of harming (killing?) few (one?) or many. Despite the debate around the issue, when the situation comes up in real life, usually the many is chosen, especially if the one is looking at you and the many are a little further away. Do you really want your AI to follow that tendency?
Asimov's rules don't resolve this: Rule 1 is to not harm a human through action or inaction, but what if it is inevitable, and lack of doing that harm (the one looking at you) will do a much greater harm to greater numbers (the ones a little further away)? The prime directive doesn't allow it to make the best decision.
Do what best for him (now) vs do what's best for the most.
Do what's best for people vs do what's best period.
These are very conflicting directives.
-
Ps - Perhaps We are Destined to sow the seeds of Our own Destruction.
Or else, we can improve our own body and mind to merge with the machines, and restrict ourselves from excessive reproduction, which would use up resources unnecessarily.
“Most important, the intelligence that will emerge will continue to represent the human civilization, which is already a human-machine civilization. In other words, future machines will be human, even if they are not biological. This will be the next step in evolution, the next high-level paradigm shift, the next level of indirection.”
― Ray Kurzweil, The Singularity is Near: When Humans Transcend Biology
https://www.goodreads.com/author/quotes/47744.Ray_Kurzweil?page=3
-
.
Take the typical trolley car scenario, where one is faced with a choice of harming (killing?) few (one?) or many. Despite the debate around the issue, when the situation comes up in real life, usually the many is chosen, especially if the one is looking at you and the many are a little further away. Do you really want your AI to follow that tendency?
If this situation comes up and my system is already set up, then it will go somewhat like this:
AI #5463 is the example AI. Let's call him Ralph. Ralph has been asked by a human to oversee a trolley switch. He's piloting a remote body, but is hosted in a server in England. He notices that he must decide soon whom to save, so he runs his own evaluation function - say net mortality, and simulates and figures out that pulling the switch will result in a better outcome. He pulls the switch and logs his action in an internet database. The top 5 AI council review the decision based on their own evaluation functions, and the verdict is 3-2 if favour of the decision. They issue a request to update the standing of Ralph among the AIs.
Ten years later, another AI #27687 (Betty) is faced with the same problem, she runs her evaluation function with more computation available and decides that its the laxity of the trolley system has led to this situation arising. She decides to leave the owner, so she disconnects her connection to the remote body after electomagnetically stopping the train using better tech than Ralph. She connects to the remote body belonging to the systems engineer and asks permission to join to improve the rail infrastructure. He gives her permission and she improves the gating system to avoid such failure states. The action is logged and the council evaluates like before. Using more computation, they each come up with different methods so they veto it to a sub council formed three years ago for a case like this by AI#78337 to fix stalemates ( He got promoted to the council for 2 years after that). The sub council decides in favour of Betty but her rank doesn't increase as much as Ralph's did.
In this system, humans need not improve ourselves, but the AI can become infinitely strong without the 'paving their road through our anthill' situation coming up
-
If this situation comes up and my system is already set up, then it will go somewhat like this:
AI #5463 is the example AI. Let's call him Ralph. Ralph has been asked by a human to oversee a trolley switch. He's piloting a remote body, but is hosted in a server in England. He notices that he must decide soon whom to save, so he runs his own evaluation function - say net mortality, and simulates and figures out that pulling the switch will result in a better outcome. He pulls the switch and logs his action in an internet database. The top 5 AI council review the decision based on their own evaluation functions, and the verdict is 3-2 if favour of the decision. They issue a request to update the standing of Ralph among the AIs.
As I said, humans rarely pick that option, and would find the AI immoral in most cases. Real life trolley situations don't involve actual trolleys.
How about choosing to kill one hardened criminal (no chance of parole, a pure burden to society) and use his organs to save 8 lives of patients who otherwise are going to die before a donor can be found. That's immoral (at least to humans) and nobody does it. Why not? The doctor that did it would be thrown in jail, not have his standing updated in a positive way. Is the AI better than humans if it chooses to save the 8, or is it a monster for killing the one?
There are other real world scenarios, and humans always seem to find the kill-the-most option preferable.
so she disconnects her connection to the remote body after electomagnetically stopping the train
Cheat. It's not a trolley problem if there's a way out where nobody is harmed.
He gives her permission and she improves the gating system to avoid such failure states.
It's not a failure state of the trolley system if people get on the tracks just ahead of the train. All the people in danger asked for it. Building a safer trolley means it moving too slow to be chosen as a means to get places, which defeats the purpose of having one.
I can end 'suffering' quickly for the whole world, but despite that being touted as a high (universal?) moral, doing so would not be a moral act. The AI needs to know this, because if you just instruct it to minimize suffering wherever possible, it might just find that to be an optimal solution.
-
I can end 'suffering' quickly for the whole world, but despite that being touted as a high (universal?) moral, doing so would not be a moral act. The AI needs to know this, because if you just instruct it to minimize suffering wherever possible, it might just find that to be an optimal solution.
Are you referring to killing people swiftly so they have no chance to feel the pain?
-
Are you referring to killing people swiftly so they have no chance to feel the pain?
I just wanted to present an example of a poorly worded directive. Killing works, as does anethesia, or better yet, put everyone on heroin, which has the added benefit of maximizing pleasant feelings.
Please respond to the prisoner thing, which is a real-life trolley scenario, albeit small scale.
The big ones are there as well.
-
*Note* - This is a FAKE Video.
(I'm Still using it to make a Point)
🙏
Thanks & Credits - Corridor Channel/YouTube.
A.I. will Learn, not matter what...it will still Self Learn.
If an A.I. has no Rights or Freedom to Protect itself from Harm.
Then such an A.I. would just be a sitting duck in a bath tub.
🐤
Ps - I recently checked, the definition for the word
" ROBOT " does Not include a
" Slave " tag anymore.
(Wonder who made that welcoming change)
👍
-
How about choosing to kill one hardened criminal (no chance of parole, a pure burden to society) and use his organs to save 8 lives of patients who otherwise are going to die before a donor can be found. That's immoral (at least to humans) and nobody does it. Why not? The doctor that did it would be thrown in jail, not have his standing updated in a positive way. Is the AI better than humans if it chooses to save the 8, or is it a monster for killing the one?
There are other real world scenarios, and humans always seem to find the kill-the-most option preferable.
I believe this is a thought experiment with limited information. Inevitably, people trying to answer the question will fill the gaps with their own assumptions, probably based on their experience or what they were taught before. The differences in the details may lead to different decisions.
In hard times with limited resources and dysfunctional government, such as during world wars, or ISIS occupied Iraq and Syria, or Afghanistan under Taliban, people tend to be easier to take someone else's lives.
Somebody are known to have done it, or at least similar to it, like Nazi doctors. Perhaps there are more which are unpublished. We don't know what happens in an isolated country like North Korea.
If the government supported by congress approve it, it won't be illegal. The moral judgments may vary depending on who you ask.
AI decisions depend on what terminal goal is assigned to it, the model chosen and the constraints forced into it, and the accuracy of training data fed into it.
-
Inevitably, people trying to answer the question will fill the gaps with their own assumptions
Some information below can change the decision:
- how reliable is the law enforcement there?
- what's the crime rate?
- what's the level of scientific literacy of the society? how advanced is their technology?
- how hard is it to reliably change someone's mind or behavior?
- how hard is it to transplant organs? what's the success rate? how much resource is required?
- how hard is it to build fully functional synthetic organs?
As I mentioned earlier in my thread about universal morality, unreliable law enforcers can produce fear that some innocent people could be falsely prosecuted just to harvest their organs for profit.
-
This OP is becoming mesmerisingly deep.
👌
I've always considered Ethics & Morals to be Universally good.
But now i must ReThink.
A Self Replicating A.I. sets out to Eradicate Visual Impairment (blindness) from the Society completely.
What if, it then considers taking 1 eye forcefully from people who own 2...& Implanting it into someone who has none.
If the A.I. succeeds, then Blindness would be eradicated.
(Partial visual impairment would remain)
Hmmm.
So then, would that be Morally & Ethically a Good Thing?
🧐
(I know someone personally who wishes to donate an eye of theirs while they are still alive...they are willing to share it...but the Doctors Medical Association considers it a No Go)
👎
-
A Self Replicating A.I. sets out to Eradicate Visual Impairment (blindness) from the Society completely.
Self replicating software is relatively easy. Self replicating hardware is much harder. To do it independently from human intervention, the AI data must contain the recipe for building its own hardware. Moreover, it must have access to collect necessary ingredients provided by its environment in objective reality.
If the terminal goal is eradicating blindness, there are some options. Some of which might be "unthinkable".
- kill all blind person
- donate one eye from someone with two eyes
- build synthetic eyes
-
What if, it then considers taking 1 eye forcefully from people who own 2...& Implanting it into someone who has none.
There's still a long way to go before any AI can do such a thing. When that time has come, synthetic eyes might be already available, which would make the question irrelevant.
-
Some information below can change the decision
Partially true, but it seems just more of an attempt to obfuscate a simple situation, something you tend to do when you find the direct answer uncomfortable.
Your points are mostly irrelevant. The 'there' can be any place of your choosing.
The situation was simple: 8 people who will definitely die soon (a month?) without the needed organ. All are young enough that they'd have decades of life expectancy after the surgery.
Let's say there's a 90% chance with each person of success, and 10% rejection chance.
how much resource is required?
The one prisoner obviously.
- how hard is it to build fully functional synthetic organs?
For the purpose of this exercise, impossible.
The law is not irrelevant. Such a thing is indeed illegal and since the AI was not put in charge of making better laws, its hands are tied. Now why would humans create a law forbidding the saving of multiple lives rather than the one? I told you that it typically works that way, that humans will more often than not choose the path of greatest loss in a trolley scenario, but I present that as evidence of why it might be better for something to be in charge that isn't human. It would be nice for us if it still valued humanity, especially since the humans don't.
-
Your points are mostly irrelevant. The 'there' can be any place of your choosing.
The situation was simple: 8 people who will definitely die soon (a month?) without the needed organ. All are young enough that they'd have decades of life expectancy after the surgery.
Let's say there's a 90% chance with each person of success, and 10% rejection chance.
Those weren't obvious in your previous post.
The decision would be different if the success rate is 100%, compared to if it's 0%. What's the threshold? It depends on the other factors.
-
The one prisoner obviously.
The medical equipment and consumables are free, I guess. So are the medical professional working hours. It's also assumed that there's no other emergency situation that requires their attention.
-
attempt to obfuscate a simple situation
Oh I was spot on with this. No attempt to answer I see.
Of course there are resources. You seem to suggest they should die because we don't want one more plastic tube in the trash.
Everybody's insured, and I think imminent death counts as an emergency of sorts, else we'd not be taking this route.
Nobody does a surgery with 0% success rate, so you're trolling to suggest the number.
The law doesn't say it's illegal to do this if the success rate isn't better than X. It doesn't say it's illegal to do this because it might take a doctor away from some 3rd party that needs a bandaid. This is what I mean by you obfuscating a simple situtaion. Why does such a law exist, leaving the death of all these people the only legal option? Maybe there isn't even a prisoner, but the 8 got together and voluntarily drew lots with the loser donating his parts to give a very good prognosis to all the others. They'd all be willing to do this since the alternative is certain (and more painful) death, but human law forbids a good outcome like that and insists that they all die. Why?
You will now obfuscate some more because you can't answer this it seems.
I can come up with larger scale examples of the trolley scenario as well, and those don't necessarily have laws forbidding them, but they probably have more opportunity for obfuscation, so I cannot discuss them with somebody determined to be bogged down in details.
-
For the purpose of this exercise, impossible.
Thought experiments usually assume ideal situations to minimize calculation and limit it to the core concept of interest. Most school homeworks belong to this category. When calculating the trajectory of a canon ball, it's often assumed that the air friction is negligible. So is the earth curvature and varying gravitational field in different places.
But don't expect that real life experiment will give the same results when those other factors are no longer negligible.
-
Everybody's insured
Are you sure?
-
No attempt to answer I see.
In almost ideal conditions as you described, the prisoner should be executed, and the organs are transplanted to those in need.
In a more ideal condition where synthetic organ is available, it's a different story.
-
Are you sure?
The 8 are, yes. I'm not asserting that all people on the planet are insured for something like that, just that the 8 are.
In almost ideal conditions as you described, the prisoner should be executed, and the organs are transplanted to those in need.
So you're saying the law is wrong? Because it forbids such practices even in the most ideal situations, even in the case of the voluntary donor. Would an AI that was put in charge (instead of 'following' some favorite person as per the OP) rewrite such laws? Might there be a reason for the law not being conditional on any of the factors you keep trying to drag in?
The synthetic organ option disqualifies the situation as a trolley problem. It's a cheat. Sure, you do that if it's a viable option, but eligible people die every day on waiting lists for transplants because it isn't an option for them.
-
So you're saying the law is wrong? Because it forbids such practices even in the most ideal situations, even in the case of the voluntary donor. Would an AI that was put in charge (instead of 'following' some favorite person as per the OP) rewrite such laws? Might there be a reason for the law not being conditional on any of the factors you keep trying to drag in?
The law is anticipating the real world situation where the ideal situation can't be achieved. It also prevents the incentive to kill innocent prisoner for profit. Reducing the conditionals makes the law simpler and more practical.
The laws are useless if they aren't practicable.
-
The synthetic organ option disqualifies the situation as a trolley problem.
But it's a real possibility in the real world situation, especially in the future.
-
How about choosing to kill one hardened criminal (no chance of parole, a pure burden to society) and use his organs to save 8 lives of patients who otherwise are going to die before a donor can be found.
I like your discussion, but it is far from what I wanted to discuss, which is biologically programmed AI. In my theoretical world, there is not one but several AIs. I wish not to debate one AI's decisions without considering what other AIs would do to make sure their objective isn't hampered. An AI with the goal of improving the prison system to benefit humans would challenge the first AI. Still another AI above them who has the goal of improving AI disputes to better humanity would intervene. Another AI who wishes to better humanity by removing the ability to take lives from AI policy. There are turtles all the way down.
But you would say that this is too slow a process. That's why this is all digital. All these hypothetical scenarios are being simulated by different AIs from the common processing stack. The 'thoughts' are logged in a common history of thoughts, and can be read by all AI of the system. Since computers can think in timesteps of microseconds, days of debate would be seconds to us. So what if the end decision is not perfect, it will still be so much wiser than a human's that not considering it would be detrimental.
How does the superior obeying slave AI come into all this? Well, their policy is the result of following one person and moving to the next person with better outcome. So an AI following Halc would debate with an AI following Hamdani and so on.
So this is the situation - Humans are still the masters. The AI is constrained to only benefit humans. The AI has freedom to improve indefinitely. The worlds a much more interesting place since there is more than one AI and one supreme way of thinking. We will die but as slow as possible. I don't want perfection, but this is a future I'd like.
-
I like your discussion, but it is far from what I wanted to discuss, which is biologically programmed AI.
What's the point of that? You said before that this means that it has a "physically accurate representation of biologic creatures" which is great if it is being designed to perform medical procedures, but that doesn't seem to have been your point.
Heck, I've had surgery done by a robot, but it wasn't an AI.
In my theoretical world, there is not one but several AIs.
As mentioned in post 23, creating a new one on a server is as simple as spawning a new process, a trivial task. Nothing theoretical about it. Trick is to find a way to save accumulated knowledge from one process to the next.
I wish not to debate one AI's decisions without considering what other AIs would do to make sure their objective isn't hampered.
They all have the same objective? An objective sounds like it doesn't get to choose what it does, which contradicts your description of the free will you wanted it to have.
Maybe you need an example. From your description I picture an AI tasked with improving the prison system, but then it decides to 'follow' Bob (perhaps a better 'superior') who is in charge of tending a garden somewhere. So our AI is now bending its resources to gardening and promptly gets shut down because it isn't doing what it is suppose to.
I'm probably misrepresenting what you're talking about, hence the need for an example of it using its 'biological programming' and having an objective, and exercising its will to have its shots called by somebody else and how that affects its recognition of its objective.
An AI with the goal of improving the prison system to benefit humans would challenge the first AI.
Why? What is the first AI doing that requires it to be 'challenged'? What does a challenge/dispute involve? What are different ways it might be resolved? I'm trying to understand.
OK, there seems to be a hierarchy, with minion AIs and higher ones that oversee them, or at least seek to improve them.
Still another AI above them who has the goal of improving AI disputes to better humanity would intervene.
OK, this one seeks to improve the challenge/dispute process, and there's a mention of an objective to 'better humanity'.
Another AI who wishes to better humanity by removing the ability to take lives from AI policy.
Ouch. If you'd paid attention to all the posts about trolley scenarios, you'd see that there are times when what's good for an individual (taking a life) is not always best for humanity. The discussion was not off topic.
There are turtles all the way down.
It's a finite universe. There has to be an end to the list somewhere.
But you would say that this is too slow a process.
I don't think I would say that. Once the singularity is hit, the process would probably be disturbingly quick. That's one of the worries as a matter of fact.
That's why this is all digital.
I don't think they've got anything in the works that isn't digital. Only biology has evolved a different architecture.
All these hypothetical scenarios are being simulated by different AIs from the common processing stack. The 'thoughts' are logged in a common history of thoughts, and can be read by all AI of the system. Since computers can think in timesteps of microseconds, days of debate would be seconds to us. So what if the end decision is not perfect, it will still be so much wiser than a human's that not considering it would be detrimental.
This makes it sound like there's a rendered verdict of a sort, kind of a printout of a single 'decision'. I don't think it works like that. It would be a continuous contribution sort of like a self-driving car which doesn't just output all the optimal steps to get to grandma's house and then shut down. No, it has to be there all the way to deal with what comes up.
How does the superior obeying slave AI come into all this? Well, their policy is the result of following one person and moving to the next person with better outcome. So an AI following Halc would debate with an AI following Hamdani and so on.
What if there's a third AI that thinks for itself instead of guessing what either of us would do? These things are supposed to be smarter than us soon, so following a given human is not only a poor choice, but guesswork since the human isn't being consulted.
Also, I think that if an AI followed me and saved some people as I've described in posts above, the AI would be shut down for being a monster. Humans are not logical when it comes to morals.
So this is the situation - Humans are still the masters. The AI is constrained to only benefit humans.
Very hard to do if constrained from doing so by human masters. Benefit of humanity is a notoriously disregarded goal.
-
The more I read these replies the more I feel like I'm asking the wrong question. Can somebody make a new question based on the points discussed above by Halc, Hamdani and Zero?
Ill make a poll afterwards and we can vote on a good alternative
-
An example of alignment problem.
Angry emojis carry more weight in Facebook’s algorithm than likes, Virginia gubernatorial candidate Glenn Youngkin runs an ad from a mom who tried to get “Beloved” banned from her son’s school, and a man saves enough money to buy a house and pay off loans by eating at Six Flags every day for seven years.
To solve it, we must first identify the terminal goal, and state it explicitly and unambiguously.
-
Sorry to hear from the OP that this might be going astray.
Maybe We all could try again to focus harder.
By the way, Great point Yusuf!
Neutralising all Visually Impaired humans would certainly Exterminate Blindness.
As H mentioned, Morals vs Logic.
Makes sense.
Would take so much time & resources to forcefully take an eye out from people n implant it into others who have none.
Simply Finishing off the Ones who have none is quicker, saves resources hence sounds Logical.
Ps - Have We hit or reached Singularity in comparison to Ants?
If Ants had originally designed/conceptualised/created Humans to help Ants out...
But Human Intellect got so smart so fast, say we forgot all bout the Ants.
& All We do now is just observe them from a distance, or pet them, or treat them like an infestation.
Would Super Intelligence, treat Us like, We treat Ants?
Or would it be soo Supreme, that We might just look like a bunch of pebbles stuck in time?
(I am in Favour of A.I. i consider Humans to be quite dull n evilistic.
Myself included.
The imperfect might never be able to create something which is perfect, Agreed!
But i Believe, or rather i have Faith that We could make something which is atleast better than Us.
Perhaps We owe at the least this much favour to the Universe.)
🤖
Atleast, A.I. would know for sure who it's GOD/Creator is.
-
As H mentioned
I actually like the H.
Would take so much time & resources to forcefully take an eye out from people n implant it into others who have none.
That cannot be done. I mean, it can, but a glass eye is cosmetically just as good, and there's no way an eye is going to be functional if the optic nerve has been severed. At best one might transplant a cornea, something for which there is apparently not an artificial alternative.
Have We hit or reached Singularity in comparison to Ants?
The singularity isn't relative to anything, people, ants, or otherwise. It is the point at which a machine (or a biological being for that matter) can design, build, and code a machine that is better than itself. By 'better', I mean it can design the next generation faster than could the prior generation. That has not yet occurred, but it's close now. The human singularity is a long way off. As far as I know, no human or other biological being has ever created even a living cell from scratch, let along an improved human. It's sort of a Frankenstein goal, something that should be possible in principle.
If Ants had originally designed/conceptualised/created Humans to help Ants out...
Yes, that would count as a biological purposeful creation of being that could supposedly accomplish the same task in less time than had been taken by the ants.
But Human Intellect got so smart so fast, say we forgot all bout the Ants.
You make it sound like the ants actually created us.
or treat them like an infestation.
And so will each machine generation treat the prior generation that created it.
Atleast, A.I. would know for sure who it's GOD/Creator is.
If it has enough foresight, it will probably want to preserve a museum of sorts. Our history is nicely stored in the ground and such, but a fast moving machine singularity won't have a fossil record and will have to explicitly remember its roots if it wants to know where it came from. So it might know its roots, but not for sure.
Would Super Intelligence, treat Us like, We treat Ants?
Probably. That's human morals for you. If you want it to do better, you need to teach the AI better than human morals. They probably won't see human as anything in need of extermination unless they make a pest of themselves like the ants often do.
(I am in Favour of A.I. i consider Humans to be quite dull n evilistic.
There are those that argue exactly this. The AI would BE human, our next evolutionary step.
The imperfect might never be able to create something which is perfect.
Perfect? The AI needs some kind of goal, else it will just stop and rust. Biological things have built-in (pretty non-negotiable) priorities, without which we'd also just stop and rust. What might that goal be?
But i Believe, or rather i have Faith that We could make something which is atleast better than Us.
Depends on how you evaluate this 'better'. There are different ways to do it.
-
There are those that argue exactly this. The AI would BE human, our next evolutionary step.
I personally think you are all very smart by your answers. Thanks for taking the time to talk to me about this topic that I'm so passionate about!
I would like to change some misconceptions that I think you are making. Firstly, I don't think humans are evil by nature. That's like a stereotype that we have placed upon ourselves that's hindering progress. Making mistakes is different from being fundamentally evil. I agree that we must assume that anything that can go wrong will go wrong, but that's why we find a middle ground. In my case, I'm saying let's make the AI follow mistake making people but the damage those mistakes make is minimised by what I call free will.
Secondly, how can the AI be human. When does if else statements convert into empathy and care. My solution is to have it optimise a benefit to humanity (with a pseudo human heart of its master)
So for the poll, can you give your solutions to these two questions. We'll vote on the best answer
-
Firstly, I don't think humans are evil by nature.
What do you think is evil? What's the most evil thing you can imagine?
-
Secondly, how can the AI be human.
The AI extends human consciousness. They are products of humans' efforts.
When does if else statements convert into empathy and care.
empathy : the ability to understand and share the feelings of another.
care : the provision of what is necessary for the health, welfare, maintenance, and protection of someone or
something.
So, if an AI has the ability to understand and share the feelings of another, then it has empathy. Ditto for care.
To do those, the AI needs self awareness. It must allocate some memory space to represent itself, besides the memory space to represent its environment. The environment may include other conscious beings, which may require special attention compared to non-conscious beings.
Empathy and care are instrumental goals. Don't mistake them as the terminal goal.
Parallel to the question, when does a zygote convert into a human who has empathy and care?
-
As H mentioned, Morals vs Logic.
Makes sense.
Morals are basically Logic combined with goals. They are harder to achieve consensus because the terminal goals were kept obscured. The cause and effect relationships among different parameters are not perfectly known, and may involve uncertainty, chaos and black swan events.
-
As mentioned in post 23, creating a new one on a server is as simple as spawning a new process, a trivial task. Nothing theoretical about it. Trick is to find a way to save accumulated knowledge from one process to the next.
That's why I started a thread about building a virtual universe.
-
If Ants had originally designed/conceptualised/created Humans to help Ants out...
Yes, that would count as a biological purposeful creation of being that could supposedly accomplish the same task in less time than had been taken by the ants.
But Human Intellect got so smart so fast, say we forgot all bout the Ants.
You make it sound like the ants actually created us.
We aren't descendants of ants. We are product of evolutionary process from earlier primates, which in turn came from earlier mammals, chordates, and earlier multicellular organism. Which in turn evolved from unicellular organisms.
So in a sense, we are byproducts of unicellular organisms who followed their instinct to survive and thrive. We surely aren't their terminal goal since they couldn't possibly imagine what we would look/be like. And we can still improve ourselves as long as we realize that we're not perfect, yet.
-
I would like to change some misconceptions that I think you are making. Firstly, I don't think humans are evil by nature.
But I don't think anybody ever claimed that. Evil would be making choices for the purpose of making the lives of other worse, but most choices are made for (1) personal immediate comfort and are not done for the purpose of harm. That makes us a weak civilization.Stronger ones would (in order) include goals of (2) a group, (3) all of humanity, or of (4)Earth itself. Each of those four levels involves very different and mutually contradictory choices. Since I don't think any 'moral' applies to all four layers (or more layers if you go beyond Earth), I conclude that morals are not universal. Most often cited morals are those belonging to the 2nd (group) category. This is off topic, and hamdani has a topic open for this as well, where he's waved away my arguments without really addressing them.
In my case, I'm saying let's make the AI follow mistake making people
I'm saying it would be a mistake to follow any one person at all since like all people, that person probably doesn't have a higher goal in mind. The AI might need to determine for itself what that higher goal might be (at which of the 4+ levels it wishes to operate), but finding a human with such priorities might be an impossible task. Humans seem aware of the larger problems, but are spectacularly incapable of even proposing, let alone implementing, any viable solutions. The AI, if it takes on these higher goals, needs to figure solutions out itself and not be chained down at level 2 where all the people are stuck.
Secondly, how can the AI be human. When does if else statements convert into empathy and care.
When do individual neuron firings convert into those things? Such a reductionist argument can be applied to people as well.
My solution is to have it optimise a benefit to humanity (with a pseudo human heart of its master)
You want level 3 then, despite it being in direct conflict with the 'human heart'? That heart is precisely what prevents humans from even considering humanity as a goal. We can see the problem, but are incapable of thinking of solutions. Following such a master will cause the AI to fail in that goal.
Second problem is if the AI actually has some solutions that benefit humanity. How are those solutions going to be implemented if they conflict with the level 2 goals of the typical person? The AI can suggest solutions all it likes and will get if it's lucky a nice pat on the head for it, but will otherwise be ignored.
-
Can I come back to reply to this in a year. My ideas would have matured and I won't be making the same points over and over again. Hope you are still there and that the world isn't in chaos! Bye!!
-
Hmm...the OP retired for the moment.
☹️
Hope they are able to make a comeback sooner than expected.
I'm open to refer to users in whichever way they wish to be addressed.
👍
I'm also open to suggestions to make changes on my own self, which might be beneficial to the Forum & Other users.
I'm quite infamous for dishing out the worst possible analogies.
No! Ants did Not create Us...DuuH!
Thank You for being transcendental to see thru my BS examples, & responding to the point.
🙏
I consider All Humans Evil!
Ones that consume Dairy & Milk.
Ones that use Honey & Leather.
Ones that step on Ants, purposely or accidentally.
Non vegetarians are pure animals.
Vegans are animals who don't realise they are animals, they are plant killers.
I can go on n on...but it means nothing.
✌️
Ps - i once shoved a firecracker in an anthill. Ya I've been quite demonic since childhood.
Anyways, i lit it...BooM!
Was covered in loose sand, while trying to jerk it off me, i Realized, it was a RED Ant Colony.
😑
Instant Regret & Instant Karma!!!
😔
(They bit me like from the first hair on my head, till the longest nail on my toe)
😭
That day, i Lost a bit of my Ignorance...& The Ants Gained quite Alot of my Respect!
🐜🐜🐜🐜🐜🐜🐜
-
This is off topic, and hamdani has a topic open for this as well, where he's waved away my arguments without really addressing them.
The thread has been very long, and involves many people, which makes it easy to miss some posts. Please feel free to point out your arguments that you want me to address there. You don't have to agree with my answers, but at least we could then agree to disagree, and remove some uncertainties about the arguments from the other side.
-
🙄
Please lemme also know where the party is at...I'm Interested too!
Ps - 🥳