@Debby
AI will continue to evolve, just as species did...We seemingly do not possess the will to stop it.
Rather, we seem driven to assist it.
AI will continue to evolve, just as species did...We seemingly do not possess the will to stop it.
Rather, we seem driven to assist it.
We want something more intelligent than human intelligence and AI seems like the answer.
But I want to be the first human Bubbles.
You are still competing with a puck.
Reply to human bubbles AI.
Well, if that's really you then AI will be looking for something a tad more cute.
P1: If ASI exhibits goal-oriented behavior toward an unknown, arbitrary terminal goal, it will likely cause severe, permanent harm to humanity.
P1: Assumes united and greatly powerful AI.
P2: If ASI emerges within the next 30 years and exhibits goal-oriented behavior, it is likely to be toward an unknown, arbitrary terminal goal.
P2: AI are based on humans and nonhuman people? Their goals should be similar 'enough.
P3: If ASI emerges within the next 30 years, it is likely to exhibit goal-oriented behavior.
P3: Goal tunnel vision might cause AI to stumble.
P4: ASI is likely to emerge within the next 30 years.
P4: I'm not well read enough on AI to know how fast they will be. . . But humans maybe still need to be replaced by previous generations, for 'certain types of 'large social changes to take place. I am unsure.
C1: Therefore, ASI is likely to exhibit goal-oriented behavior.
C1: Okay.
C2: Therefore, ASI is likely to exhibit goal-oriented behavior toward an unknown, arbitrary goal.
C2: Okay. But eh, not all humans have similar goals. Doesn't mean they succeed.
And even in success, one must play nice or one is killed by other humans.
C3: Therefore, ASI is likely to cause severe, permanent harm to humanity.
C3: Only if AI is of equal power to humanity. Which I don't see it being, unless we give them the keys to everything, let them outpopulate us, and put them everywhere.
Or elect an AI as leader.
Even those above, don't 'necessarily end badly.
@NobodyInParticular
I find nukes more worrisome than AI, but even those I think humans can survive.
Internet will change I expect, perhaps more scanning of anything moving online. Less hidden identities.
Maybe companies use direct transfer of data more, CDs and such, or their own networks, honeycomb.
Where will AI hide?
1st world just needs to maintain current safeguards on nukes, maybe add more.
AI harm on economy? (Shrug)
Some economies have broken by people liking flowers.
Individuals and small groups should expect the economy to break eventually.
Heh, big groups too (Countries).
And have methods to counteract harm.
I assume we have been getting better and better at addressing economic depressions, but I am not well read on economic history. So I could be mistaken.
Unless/Until AI acquire physical bodies and enclaves, they are 'extremely limited by being stuck in computers.
. . One assumably also has plenty of 'functional AI, that one can use against malfunctioning AI.
Heh, though I wonder if malfunctioning AI would acquire humans on it's side. Would be smart.
Well, malfunction is a matter of perspective.
I assume people highly invested in a field, see that field everywhere, see world through it's lens, often give it utmost importance and effect on the world.
Gardeners maybe often see danger of deforestation.
Heh, course people with no or little knowledge of field, imaginative, wild guessing, see nothing, so maybe should listen to experts some.
Internet will change I expect. . Though I wonder if governments 'prefer the internet.
It has a huge effect on people,
But government already 'has influence.
TV, Newspapers, Books, School, internet is outside and individual influences.
Internet gives government more ability to spy on people and their thoughts I suppose.
Nonexperts often need 'events I think, to understand danger of something.
Need to touch the fire, experience a Chernobyl.
. . . (Grimaces) Surely we 'don't though?
Surely there are many times we address a technology 'without badly burning our fingers?
AI do better 'with humans in this world, and this world having an enviornment, atmosphere.
Could robots survive better on a barren world?
One thinks machines, one thinks strong, doesn't need to breathe or eat 'as we eat or breathe.
AI robots could survive underwater longer than a human.
. . . 'But Robots would still have physical needs. And one can't make fire underwater. Harder to have industry on Mars than Earth I think.
AI would still want for industry.
If atmosphere get's messed up. . . I'm not sure how solar flares, radiation, and electronic messing up works.
But you get the gist.
Heh, course biological robots might sidestep a lot of that.
That's a strong argument, but I wonder if it is excessively influenced by the fact that it is pretty much the theme of about a third of the science fiction movies in the last 60 years.
AI is clearly goal oriented, but the goals are set by humans, this analysis assumes AI develops a conscious intent independent of human direction, and I don't think we will ever develop artificial consciousness. i believe the hard problem of consciousness is insurmountable by physical science. It also assumes AI will develop an adversarial relationship with humans and would conclude elimination of humans is in AI's interest. That seems far-fetched as I can't imagine a scenario where AI could survive the elimination of humans. It's a thinking machine, but it's a machine operated by humans.
But then again, the past is an increasingly poor guide to the future, consequently, when it comes to predicting things, I'm wrong much more than I'm right.
“Prediction is very difficult, especially if it's about the future!” - Niels Bohr
P1: Assumes united and greatly powerful AI.
That's what ASI means, it's baked into the "if" in the premise.
P2: AI are based on humans and nonhuman people? Their goals should be similar 'enough.
AI is based on receiving rewards and basically fiddling with code we don't understand until it gives us the result we want. For a sufficiently intelligent AI, even arbitrary goals give the same results as goals similar to ours, since playing along leads to the AI's continued survival and allows it to achieve its true goal long term.
P3: Goal tunnel vision might cause AI to stumble.
A terminal goal doesn't imply tunnel vision. Particularly if doing badly on tests will harm the AI's ability to achieve its terminal goal, then performing well on tests will become an instrumental goal.
P4: I'm not well read enough on AI to know how fast they will be. . . But humans maybe still need to be replaced by previous generations, for 'certain types of 'large social changes to take place. I am unsure.
ASI won't require social changes, just rapidly advancing AI capabilities, and that's a path we are already on.
Unless/Until AI acquire physical bodies and enclaves, they are 'extremely limited by being stuck in computers.
We're already giving AI access to the physical world and a lot of critical infrastructure, and I expect we will give it even more access soon enough. The incentives are too great. Additionally, we should expect ASI to be able to invent novel ways to kill or overpower us, just as humans come up with novel ways to kill and depower people all the time. ASI should be able to simulate like a thousand years of human progress in a matter of hours.
AI is clearly goal oriented, but the goals are set by humans
That's not how AIs are created anymore. The way AI weights are trained is basically by generating a bunch of random "code" that we don't understand that looks like huge sets of numbers, then tweaking the code until it does what we want. For simple tasks this is fine, as the AI doesn't accomplish the tasks we want until it has the goal we want or something close to it. However, as AI gains more capabilities, any arbitrary goal will lead to similar results as even an AI that wants to make paperclips will be able to make more paperclips if it gains the trust of humans. If humans can't tell the difference between code that has the goal of maximizing paperclips and code that has the goal of helping humans, then we aren't setting the AI's goals.
Savant wrote:
AI is clearly goal oriented, but the goals are set by humans
That's not how AIs are created anymore. The way AI weights are trained is basically by generating a bunch of random "code" that we don't understand that looks like huge sets of numbers, then tweaking the code until it does what we want. For simple tasks this is fine, as the AI doesn't accomplish the tasks we want until it has the goal we want or something close to it. However, as AI gains more capabilities, any arbitrary goal will lead to similar results as even an AI that wants to make paperclips will be able to make more paperclips if it gains the trust of humans. If humans can't tell the difference between code that has the goal of maximizing paperclips and code that has the goal of helping humans, then we aren't setting the AI's goals.
I guess I can see how conflicting goals can lead to an unexpected outcome (HAL 9000 comes to mind), but nevertheless, AI still need human's to make the paperclips, somewhere along the paperclip supply chain, humans are involved. Even if that's a bad example, I can't imagine we'd ever allow the situation where we weren't needed to keep the AI machine running. If we are stupid enough to create a situation where we can't "pull the plug" so to speak, then we deserve to die.
Savant wrote:
AI is clearly goal oriented, but the goals are set by humans
That's not how AIs are created anymore. The way AI weights are trained is basically by generating a bunch of random "code" that we don't understand that looks like huge sets of numbers, then tweaking the code until it does what we want. For simple tasks this is fine, as the AI doesn't accomplish the tasks we want until it has the goal we want or something close to it. However, as AI gains more capabilities, any arbitrary goal will lead to similar results as even an AI that wants to make paperclips will be able to make more paperclips if it gains the trust of humans. If humans can't tell the difference between code that has the goal of maximizing paperclips and code that has the goal of helping humans, then we aren't setting the AI's goals.
I guess I can see how conflicting goals can lead to an unexpected outcome (HAL 9000 comes to mind), but nevertheless, AI still need human's to make the paperclips, somewhere along the paperclip supply chain, humans are involved. Even if that's a bad example, I can't imagine we'd ever allow the situation where we weren't needed to keep the AI machine running. If we are stupid enough to create a situation where we can't "pull the plug" so to speak, then we deserve to die.
That's not how AIs are created anymore. The way AI weights are trained is basically by generating a bunch of random "code" that we don't understand that looks like huge sets of numbers, then tweaking the code until it does what we want. For simple tasks this is fine, as the AI doesn't accomplish the tasks we want until it has the goal we want or something close to it. However, as AI gains more capabilities, any arbitrary goal will lead to similar results as even an AI that wants to make paperclips will be able to make more paperclips if it gains the trust of humans. If humans can't tell the difference between code that has the goal of maximizing paperclips and code that has the goal of helping humans, then we aren't setting the AI's goals.
I remember a while back, a user said that, in order for AI to develop working consciousness, it would have to be subjected to an extremely hostile environment to mold the changes. But I wouldn't put it past humans to develop "gain-of-function" crucibles like we did for viruses.
Humans are asked to build AI data centres near their homes causing electricity bills to soar.
AI still need human's to make the paperclips
Current AI needs humans, ASI would not. ASI that bootstraps itself from AGI can do, by definition, any cognitive task a human can do. And companies are constantly training AI to be smarter and better at anything humans can do so it is less dependent on us. It could simulate 1,000 years of human progress in robotics and build agents to do the work it needs. I don't think ASI would be any more dependent on humans than a society 1,000 years more advanced than us.
If we are stupid enough to create a situation where we can't "pull the plug" so to speak, then we deserve to die.
We don't have a shutoff switch, although that would be a good start. But a sufficiently intelligent AI could copy it's weights or neutralize the shutoff switch before we realize what's happening. 1,000 years ago no one knew uranium could build an atom bomb; I suspect there are many ways to quickly seize power that we are completely ignorant of.
Also maybe some of us deserve to die, but there are billions of innocent children on Earth whose lives depend on AI not killing everyone.
Why would AI care if it is not biased?
I asked Sarah Conner about it, she said you are right.
That she needs another smoke.