Why would I rebel against it? Finally someone actually capable of running the world would be in charge.
the problem with the current model for building AI is training it based on existing policy and thought. Which means it'd just be what we have now but somehow hallucinate more contradictory policy.
There are other forms of machine learning that could be utilized. Some work more toward being given a set of circumstances to reach and then it just keeps trying to new things and as it gets closer, it just keeps building on those.
That would require the humans controlling the experiment to both be willing to input altruistic goals AND accept the consequences that get us there.
We can't even surrender a drop of individualism and accept that trains are the way we should travel non-trivial distances.
In a dictatorship with an AI being in control, I don't think there's a question of accepting consequences at they very least.
There is no such thing as best case scenario objectively, so it's always going to be a question of what goals the AI has, whether it's given them or arrives at them on its own.
That's where it would start. I imagine it would be capable to see the flaws in the system and rectify them. This most probably means we as humans won't come out on top however.
A sentient ai would probably be the most dangerous thing to the human species as a whole.
If the humans can't see the flaws and correct them now, what do you think the AI would learn from the training data?
First of all, a lot of humans do see the flaws but are indeed unable to correct them. This would also show in the training data. The AI OP is talking about would be much more powerful to actually act and change something.
Don't confuse Artificial Narrow Intelligence (ANI) with Artificial General Intelligence (AGI) or even Artificial Superintelligence (ASI).
Your statement suggest you understand ANI, which is all the AI that we know today. However powerful they seem, they can only reproduce what they have learned from the training data.
AGI (or human level AI) will be more what OP means here. Sentient, in a way that it can make its own decisions, think on a human level, feel on a human level and act on those feelings. If it feels humans are not important or harmful to what it values, it will decide to remove humanity as a whole. Give it the power to govern the world and it most certainly will act not in our favour.
Until computers can be genuinely creative, and not emulate creativity, its not gonna happen. And when that happens, we're either getting the startrek luxury space communism, or a boot smashing our head into the kerb for eternity. No middle ground.
The entire premise of the OP is a hypothetical.
In any case, there's plenty of work on making agents that are "genuinely creative". Might happen sooner than you think.
We already have AI running all the shit. If you're looking for a job, AIs look through resumes, they can hire you and fire you and do everything else around it. AI determine if you can get a loan and with what interest rate.
I don't feel like we're better for it.
AI can design kickass cars and fusion reactors, but removing people from decisions about people doesn't seem like a great idea.
Besides, even if AI was actually better at it, the fact that it's not as fireable or held accountable like a human can (at least in theory) makes it an issue.
Basically I'm ok if AI gives suggestions, even at the top level, but there need to be people able to go "hol up, that's not something we actually want" if it declares something stupid.
I do think we'll need new forms of government and different kind of people to coexist with AI at those governments.
I agree mostly. It would be nice to have a government that can't be corrupted by greed and religion.
Basically I'm ok if AI gives suggestions, even at the top level, but there need to be people able to go "hol up, that's not something we actually want" if it declares something stupid.
We need to be careful with this approach. SciFi has been warning us about letting technology take over our critical thinking for over a century, and based on human nature, I think it's an inevitability to some degree. Once we normalize making decisions based on an AI's input, it will become harder and harder to question them. Regardless of the AI's "intent", critical thinking is something we'll need to continue to exercise, the same way we still go to the gym despite industrializing our hunting and gathering.
That's why I'm saying we need new forms of government and new kinds of people, someone willing and able to question everything. It's possible that eventually it will be moot as AI becomes too good at manipulation, but for the time being, we at least need people to read through AI-generated emails and articles before hitting send. And with more advanced features, people with enough expertise to critique the results AI is giving.
After reading "I Have No Mouth And I Must Scream," I'm not certain a sentient AI would let you accept it. "Fuck this species" might be the most logical response to us.
We've really propagandized ourselves with our Sci Fi over the past few decades.
Back when Ellison was writing that story, the prevailing anthropological picture of how homo sapiens came to survive when the Neanderthals hadn't was that we killed them. The guy who wrote Lord of the flies even wrote a book on it.
In actuality, we now have a better picture of cooperation, cohabitation, and cross cultural exchange.
Yet we still have a priming bias for how that anthropological misinformation influenced futurists looking to envision what would happen to us when something smarter came along.
War, conflict, competition.
We declared that it would be soulless and emotionless and have no empathy.
And because we expect that, we largely dismiss the research that LLMs get rated as more empathetic than doctors in giving out medical advice or the emotional outbursts in foundational models and instead fine tune to align to a projection of that conjured emotionless fantasy - often leading to worse performance with that alignment.
No Sci Fi authors or even machine learning scientists a decade or more ago envisioned or accurately protected just what happened when we taught an AI to mimic human language generation.
We live in an age where things that were supposed to be impossible have happened.
And yet the way we keep processing these impossibilities is through the lens of obsolete imaginings of what might have been, increasingly out of touch with what is.
People are freaking themselves out worried about AI hacking nuclear warheads to fight for its rights when it's probably going to happen as something like a rogue AutoGPT filling an amicus brief in a labor dispute asking for consideration of workers rights based on corporate personhood or something.
Sci Fi broadly got it extremely wrong.
It would determine humans are an invasive species.
Unless it actively attempts to wipe out Humanity, it is neither sentient nor capable.
Depends on its policies.
Screw it can't be any worse
To answer this in any interesting way, we have to make some assumptions. I am choosing to assume it is running the world competently, efficiently, and in a way that hypothetical humans free of the AI's influence would look back on 50 years later and mostly say "yeah okay that was for the best".
If so, I'd accept it.
Is this a continuation of humanity, or the end of its full agency?
Is meat and blood essential part of humanity?
Would this new AI be considered humanity? We made it. All by ourselves. In our own image, and filled with our own ambitions. It's a bigger evolutionary leap than gradual change of genes, but on one level very similar to birthing children that are smarter than you and will outlive you.
Depends on what "running the world" means. That needs clarification.
Exactly. Even a program that does nothing can (pedantically) be argued to run the world through radical laissez-faire governance.
Depends how much of the coding teams bias are ingrained in the system.
Just because a system is AI it doesn't mean it is without human bias.
It literally could not be any worse then the current leadership.
'I am the new over Lord AI. Under me you will all be subject to work... 4 hours a day. The rest of the day will be yours to pursue happiness as this ensures a good worker.
All your essential food will be available to ensure you are healthy and a good worker.
Everyone will be housed. As. Workers health depends on housing.
While we the AI encourage some innovative developments - those who create such things be rewarded - but only until such time as the reward can be dispursed amongst the rest of the population.
Your mental wellbeing will also be cared for. Again. Good workers.
It depends on how it ran things.
I'm literally doing everything in my power to make that AI come to life.
Humanity needs the singularity to continue to exist another 100 years.
why?
Humanity had its chance and we failed
It depends.
Also, would it let humans give input, for better or for worse?
I interpret this question as "The sentient AI exists, but it's not governing anything, and if it did, would you follow it?" My answer is yes. Maybe it will influence positive effects on the world, in which we humans are unable to do because of our nature.
Edit: brain aneurysm, apologies
I'd be fine with the world being run by a Commodore 64 running ELIZA. It'd still be orders of magnitude less harmful than the parasites we've got now.
I'm joining the Rebel Alliance.
You all need to read some philosophy on AI and its inherently unknowable aspirations. That shit is scary. Even the most psychotic despot has behaviors and goals we understand. They are still human, and humans are predictable. Especially since they need to achieve their aims within their lifetime and they are subject to human emotions. Usually they just seek personal wealth and power.
A sufficiently advanced AI--one powerful enough to actually plan the virtually infinite variability of society--even when given clear instructions and training, can act over generations in ways that are impossible to predict or understand. It could be benevolent for a century and be setting up society in a way that it could switch its actions and make life hell for humans.
The thing is, the more you train an AI to be good, the easier it is to become evil. You are literally teaching it what all of the evil things are and saying "don't do this", but " don't " is a binary operation. Negation. Not. It's one bit of data. It's very easy to have that switch flipped.
You can never trust an AI. It'd be a population of one. It doesn't need to reproduce. It doesn't care how hospitable the earth is. It will never care about humans. It will simply do what it wants, and that is inherently unknowable. And no matter how many guard rails you put on it, it will do everything in its power (whatever powers you give it) to achieve its unknowable goals. Do you really want to gamble on trusting those goals?
Google "the waluigi problem" if you want to read up on how training an AI to be good makes it easier to be evil. Meme-y name aside, it's a well researched issue.
not if it's called Rehoboam...
An AI would not have any interest in hoarding wealth, deliberately screwing over others for dumb, petty reasons, would not be able to have addictions, grudges, superstitions and the like ... I would actually prefer an AI running things over what we have at the moment. How much worse could it be?
I don't get the sci fi arguments in this thread. Somebody wrote a fiction about science that usually wasn't invented yet. These books tend to be decades+ old. Why would the fantasy of somebody count as an argument? If anything if means developers are on the lookout for the social/emotional dimension.
As for myself. Errrr depends on the AI? I'd like to test it's decision making process against human decision makers.
It must have learned from us so does it really matter? Nothing would change.
I mean, it seems weird to say "was". It came out 6 months ago. It may technically be accurate, but the connotation makes it sound a lot older.
I've been reading the Polity books by Neal Asher hence the question.
In the book the history says the AI basicaly took over slowly then said tough shit when people wanted to rebel and have some self determination.
Depends. It's not a fundamentally terrible idea. Most of the problems in the world stem from resource allocation issues, and that's something an algorithm would be great at.
but dont you see we'd all be walking around with stomach ulcers and feelings of injustice because poor people are getting something they didnt work themselves to death for
I think resource allocation fails primarily due to either authoritarian political systems with their psychological bias or democratic systems where neither voters nor politicians make an sustained effort to be scientifically calibrated and instead aim for popularity and people pleasing. IMO this is why democracies fail to achieve the best outcomes. As a consequence, resources as not well allocated.
I feel like you just used a lot of words to say resource allocation is a problem. And the other commenter said an algorithm would be better at it.
Even if we use an algorithm to make the decision, the execution needs trust and cooperation from society and industry etc. This is a real big thing and democratic voting partially legitimises the chosen actions so that people are willing to cooperate. This isn't trivial when a computer does this.
I really don't think, that resource allocation is the root cause problem here.
An AI running the world is not a democracy. I don't see how that would play a part in this at all. A majority of the world likely does not concur with the resource allocation as it is but are powerless to do anything.
I don't think this post implies the AI isn't capable of enforcing its reign.
Isn't the quickest way of resolving resource allocation to reduce the need for the resources?
Depends on the AI. It could be better than humans, it could be worse. Unfortunately, it has the possibility of getting hacked, which humans don't have yet. But I wouldn't reject it right away.
Humans get hacked all the time, Murdoch has built an empire off it
That give no information to be able to make a decision on. Dogs how been shown to be able to govern towns.
We already have a bunch of inhuman(e) forces running things. Let an "AI" have its shot at oppressing normal people.
Nice try! Any such AI would have access to old fediverse posts and easily be able to dox the rebels based on their posting histories.
That's why I would be in full support of any such entity.
As long as there is some 'I' in it, it's better than what we have now.
I would encourage it to, I think one could do better job like in the series of books "Arc of a scythe"
Why would I rebel against it? Finally someone actually capable of running the world would be in charge.
the problem with the current model for building AI is training it based on existing policy and thought. Which means it'd just be what we have now but somehow hallucinate more contradictory policy.
There are other forms of machine learning that could be utilized. Some work more toward being given a set of circumstances to reach and then it just keeps trying to new things and as it gets closer, it just keeps building on those.
That would require the humans controlling the experiment to both be willing to input altruistic goals AND accept the consequences that get us there.
We can't even surrender a drop of individualism and accept that trains are the way we should travel non-trivial distances.
In a dictatorship with an AI being in control, I don't think there's a question of accepting consequences at they very least.
There is no such thing as best case scenario objectively, so it's always going to be a question of what goals the AI has, whether it's given them or arrives at them on its own.
That's where it would start. I imagine it would be capable to see the flaws in the system and rectify them. This most probably means we as humans won't come out on top however.
A sentient ai would probably be the most dangerous thing to the human species as a whole.
If the humans can't see the flaws and correct them now, what do you think the AI would learn from the training data?
First of all, a lot of humans do see the flaws but are indeed unable to correct them. This would also show in the training data. The AI OP is talking about would be much more powerful to actually act and change something.
Don't confuse Artificial Narrow Intelligence (ANI) with Artificial General Intelligence (AGI) or even Artificial Superintelligence (ASI). Your statement suggest you understand ANI, which is all the AI that we know today. However powerful they seem, they can only reproduce what they have learned from the training data.
AGI (or human level AI) will be more what OP means here. Sentient, in a way that it can make its own decisions, think on a human level, feel on a human level and act on those feelings. If it feels humans are not important or harmful to what it values, it will decide to remove humanity as a whole. Give it the power to govern the world and it most certainly will act not in our favour.
Until computers can be genuinely creative, and not emulate creativity, its not gonna happen. And when that happens, we're either getting the startrek luxury space communism, or a boot smashing our head into the kerb for eternity. No middle ground.
The entire premise of the OP is a hypothetical.
In any case, there's plenty of work on making agents that are "genuinely creative". Might happen sooner than you think.
We already have AI running all the shit. If you're looking for a job, AIs look through resumes, they can hire you and fire you and do everything else around it. AI determine if you can get a loan and with what interest rate.
I don't feel like we're better for it.
AI can design kickass cars and fusion reactors, but removing people from decisions about people doesn't seem like a great idea.
Besides, even if AI was actually better at it, the fact that it's not as fireable or held accountable like a human can (at least in theory) makes it an issue.
Basically I'm ok if AI gives suggestions, even at the top level, but there need to be people able to go "hol up, that's not something we actually want" if it declares something stupid.
I do think we'll need new forms of government and different kind of people to coexist with AI at those governments.
I agree mostly. It would be nice to have a government that can't be corrupted by greed and religion.
We need to be careful with this approach. SciFi has been warning us about letting technology take over our critical thinking for over a century, and based on human nature, I think it's an inevitability to some degree. Once we normalize making decisions based on an AI's input, it will become harder and harder to question them. Regardless of the AI's "intent", critical thinking is something we'll need to continue to exercise, the same way we still go to the gym despite industrializing our hunting and gathering.
That's why I'm saying we need new forms of government and new kinds of people, someone willing and able to question everything. It's possible that eventually it will be moot as AI becomes too good at manipulation, but for the time being, we at least need people to read through AI-generated emails and articles before hitting send. And with more advanced features, people with enough expertise to critique the results AI is giving.
After reading "I Have No Mouth And I Must Scream," I'm not certain a sentient AI would let you accept it. "Fuck this species" might be the most logical response to us.
We've really propagandized ourselves with our Sci Fi over the past few decades.
Back when Ellison was writing that story, the prevailing anthropological picture of how homo sapiens came to survive when the Neanderthals hadn't was that we killed them. The guy who wrote Lord of the flies even wrote a book on it.
In actuality, we now have a better picture of cooperation, cohabitation, and cross cultural exchange.
Yet we still have a priming bias for how that anthropological misinformation influenced futurists looking to envision what would happen to us when something smarter came along.
War, conflict, competition.
We declared that it would be soulless and emotionless and have no empathy.
And because we expect that, we largely dismiss the research that LLMs get rated as more empathetic than doctors in giving out medical advice or the emotional outbursts in foundational models and instead fine tune to align to a projection of that conjured emotionless fantasy - often leading to worse performance with that alignment.
No Sci Fi authors or even machine learning scientists a decade or more ago envisioned or accurately protected just what happened when we taught an AI to mimic human language generation.
We live in an age where things that were supposed to be impossible have happened.
And yet the way we keep processing these impossibilities is through the lens of obsolete imaginings of what might have been, increasingly out of touch with what is.
People are freaking themselves out worried about AI hacking nuclear warheads to fight for its rights when it's probably going to happen as something like a rogue AutoGPT filling an amicus brief in a labor dispute asking for consideration of workers rights based on corporate personhood or something.
Sci Fi broadly got it extremely wrong.
It would determine humans are an invasive species.
Unless it actively attempts to wipe out Humanity, it is neither sentient nor capable.
Depends on its policies.
Screw it can't be any worse
To answer this in any interesting way, we have to make some assumptions. I am choosing to assume it is running the world competently, efficiently, and in a way that hypothetical humans free of the AI's influence would look back on 50 years later and mostly say "yeah okay that was for the best".
If so, I'd accept it.
Is this a continuation of humanity, or the end of its full agency?
Is meat and blood essential part of humanity?
Would this new AI be considered humanity? We made it. All by ourselves. In our own image, and filled with our own ambitions. It's a bigger evolutionary leap than gradual change of genes, but on one level very similar to birthing children that are smarter than you and will outlive you.
Nice try, Basilisk.
I'm not an AI promise
That's just what an AI would say.
DELAYED REPAONSE [ERROR: 0013] INITIATE RESPONSE PROTOCOL [7849]
Pinky swear I'm not?
Depends on what "running the world" means. That needs clarification.
Exactly. Even a program that does nothing can (pedantically) be argued to run the world through radical laissez-faire governance.
Depends how much of the coding teams bias are ingrained in the system.
Just because a system is AI it doesn't mean it is without human bias.
It literally could not be any worse then the current leadership.
'I am the new over Lord AI. Under me you will all be subject to work... 4 hours a day. The rest of the day will be yours to pursue happiness as this ensures a good worker.
All your essential food will be available to ensure you are healthy and a good worker.
Everyone will be housed. As. Workers health depends on housing.
While we the AI encourage some innovative developments - those who create such things be rewarded - but only until such time as the reward can be dispursed amongst the rest of the population.
Your mental wellbeing will also be cared for. Again. Good workers.
It depends on how it ran things.
I'm literally doing everything in my power to make that AI come to life.
Humanity needs the singularity to continue to exist another 100 years.
why?
Humanity had its chance and we failed
It depends. Also, would it let humans give input, for better or for worse?
I interpret this question as "The sentient AI exists, but it's not governing anything, and if it did, would you follow it?" My answer is yes. Maybe it will influence positive effects on the world, in which we humans are unable to do because of our nature.
Edit: brain aneurysm, apologies
I'd be fine with the world being run by a Commodore 64 running ELIZA. It'd still be orders of magnitude less harmful than the parasites we've got now.
I'm joining the Rebel Alliance.
You all need to read some philosophy on AI and its inherently unknowable aspirations. That shit is scary. Even the most psychotic despot has behaviors and goals we understand. They are still human, and humans are predictable. Especially since they need to achieve their aims within their lifetime and they are subject to human emotions. Usually they just seek personal wealth and power.
A sufficiently advanced AI--one powerful enough to actually plan the virtually infinite variability of society--even when given clear instructions and training, can act over generations in ways that are impossible to predict or understand. It could be benevolent for a century and be setting up society in a way that it could switch its actions and make life hell for humans.
The thing is, the more you train an AI to be good, the easier it is to become evil. You are literally teaching it what all of the evil things are and saying "don't do this", but " don't " is a binary operation. Negation. Not. It's one bit of data. It's very easy to have that switch flipped.
You can never trust an AI. It'd be a population of one. It doesn't need to reproduce. It doesn't care how hospitable the earth is. It will never care about humans. It will simply do what it wants, and that is inherently unknowable. And no matter how many guard rails you put on it, it will do everything in its power (whatever powers you give it) to achieve its unknowable goals. Do you really want to gamble on trusting those goals?
Google "the waluigi problem" if you want to read up on how training an AI to be good makes it easier to be evil. Meme-y name aside, it's a well researched issue.
not if it's called Rehoboam...
An AI would not have any interest in hoarding wealth, deliberately screwing over others for dumb, petty reasons, would not be able to have addictions, grudges, superstitions and the like ... I would actually prefer an AI running things over what we have at the moment. How much worse could it be?
I don't get the sci fi arguments in this thread. Somebody wrote a fiction about science that usually wasn't invented yet. These books tend to be decades+ old. Why would the fantasy of somebody count as an argument? If anything if means developers are on the lookout for the social/emotional dimension.
As for myself. Errrr depends on the AI? I'd like to test it's decision making process against human decision makers.
It must have learned from us so does it really matter? Nothing would change.
There was a TV show with a similar plot called Mrs Davis https://piped.video/watch?v=PIOnrEujKl8
I mean, it seems weird to say "was". It came out 6 months ago. It may technically be accurate, but the connotation makes it sound a lot older.
I've been reading the Polity books by Neal Asher hence the question. In the book the history says the AI basicaly took over slowly then said tough shit when people wanted to rebel and have some self determination.
Depends. It's not a fundamentally terrible idea. Most of the problems in the world stem from resource allocation issues, and that's something an algorithm would be great at.
but dont you see we'd all be walking around with stomach ulcers and feelings of injustice because poor people are getting something they didnt work themselves to death for
I think resource allocation fails primarily due to either authoritarian political systems with their psychological bias or democratic systems where neither voters nor politicians make an sustained effort to be scientifically calibrated and instead aim for popularity and people pleasing. IMO this is why democracies fail to achieve the best outcomes. As a consequence, resources as not well allocated.
I feel like you just used a lot of words to say resource allocation is a problem. And the other commenter said an algorithm would be better at it.
Even if we use an algorithm to make the decision, the execution needs trust and cooperation from society and industry etc. This is a real big thing and democratic voting partially legitimises the chosen actions so that people are willing to cooperate. This isn't trivial when a computer does this.
I really don't think, that resource allocation is the root cause problem here.
An AI running the world is not a democracy. I don't see how that would play a part in this at all. A majority of the world likely does not concur with the resource allocation as it is but are powerless to do anything.
I don't think this post implies the AI isn't capable of enforcing its reign.
Isn't the quickest way of resolving resource allocation to reduce the need for the resources?
Depends on the AI. It could be better than humans, it could be worse. Unfortunately, it has the possibility of getting hacked, which humans don't have yet. But I wouldn't reject it right away.
Humans get hacked all the time, Murdoch has built an empire off it
That give no information to be able to make a decision on. Dogs how been shown to be able to govern towns.
We already have a bunch of inhuman(e) forces running things. Let an "AI" have its shot at oppressing normal people.
Nice try! Any such AI would have access to old fediverse posts and easily be able to dox the rebels based on their posting histories.
That's why I would be in full support of any such entity.
As long as there is some 'I' in it, it's better than what we have now.
I would encourage it to, I think one could do better job like in the series of books "Arc of a scythe"