AI to cause human extinction.???

Only because the threat is so low on my list of things worth waking up in a cold sweat at 3 am for.

 

Log in or Sign up to hide all adverts.

OK, but that's kind of suggesting that you're not really interested in engaging, even as Devil's Advocate. If I'm gonna discuss, I'd like my challengers to care enough to get out of bed for it.

Please Register or Log in to view the hidden image!

 

Log in or Sign up to hide all adverts.

I have been engaging - as I don't see a whole lot of other people doing, BTW, - and what I have been attempting to convey is that I really, truly do not believe this topic needs any kind of advocates.
I see the threat you're concerned about, but don't see you offering precautions or defences against it, and even if you did, I don't see what you or I could actually do about it.

I think the "experts" are freaking out over the the wrong danger. AI is a threat to an arrangement that's been an growing threat to humanity and all life on this planet for 7000+ years. They're worried about a threat to the lifestyle we should never have adopted in the first place - that's crippled and deformed and disfigured us as a species, as societies and as individuals, and that's morphed into something even more insidious and toxic in the past 500 years. Whether AI turns out to be a help or a hindrance, I'm not sure we can recover from imperial capitalism.

Fine. At 8:30, and calm enough to get my slippers on and find my way to the bathroom.

 

Log in or Sign up to hide all adverts.

I think the original statement is premature. AI at present gives answer on a computer. It has been untrustworthy in many instances. It has a very long way to go before we have AI robots . I don't believe there are any programs beyond improving AI programs, which are intrinsically dumb, but have lots of data and fast computers.

 

Clearly AI will be used in warfare and each side will have the latest version

When one side wins its version will be "battle hardened " and it may be difficult to get it to readjust to "civilian life"

It may still feel it needs to protect its "owners" and therefore its self

There may be a need for tough love as not all those it is trying to protect share the same societal ethos and some may even be seen as an "enemy within"

Sounds a bit creepy to me .Can we turn it off?

Daisy Daisy?

 

Probably not. The people who currently own and control the most important functions won't give up the advantage they have - or think they have. It's all academic.
But do keep in mind that these predictions are mostly coming from experts who have been greatly enriched by developing AI technology, and most of the alarm is over threats to the economy which allowed it to do so. Self-serving? Maybe. Hypocritical? Probably. Atonement? Maybe. Early warning? Hardly!

 

You don't think that the technology is independent of whoever or whatever created it?

Einstein was in my opinion a good man but he could as easily have been a monster since either way we are saddled with MAD ever since it became clear what e=mc^2 entailed.

AI is now in our bones and we have to learn to live with it.

 

Untrustworthy does not necessarily equate with harmless, though.

A machine that is untrustworthy (i.e. not doing what we told it to) could very well cause lots of harm. Like, say opening a containment valve at the wrong time for the wrong reason.

AI robots are not necessary to wreak havoc. Specialized AI is already controlling a lot of automated systems, from cars to stoplights to surveillance systems.

 

Maybe it's just me, but I'm struggling in this thread to understand what everyone is imagining with regards AI, that makes the AI itself the danger? To me, AI is a tool that can help us sift through vast quantities of data in a very short time, by learning. But what it can do is very much limited by what we ask / set it up to do. The AI that can beat humans at chess can't boil and egg. The AI that helped discover a potent antibiotic (against a specific virus) by whittling down 1000s of chemicals to a few hundred, it can only do specific things. This is what AI is, currently. A tool. It is the people that use it that will generate the existential threats, as far as I can see. It will be people that set AI to the task of identifying a new biological agent for use in warfare, for example. It will be people who set AI to discovering new sources of minerals, to use a previously offered example. But it will be humans who decide how to react to such things.

Now, we could build an autonomous system with AI, such as has been done in self-driving cars. But in such cases it is humans deciding and determining what agency to grant the system. At no point can the AI go outside what the humans have allowed it to do. Sure, at least not yet. AI, at least for the forseeable, is all about training a system to a very specific task, which it (hopefully) can do far better (quicker, more accurately, more efficiently etc) than us humans.

There may come a time, although there may not, where an artificial general intelligence is developed that can be put to any task, and it will learn from general experience, rather than having a specific AI for each specific task. But we're not there yet, and may never be.

There may also come a time when we achieve artificial sentience. But that, currently, is speculative fiction, at least imo.

So, the idea that AI will treat us as an impediment, and those such existential threats, are, or should be, consigned to the realm of science fiction at the moment imo. The real threat that AI poses, imo, is to allow people with nefarious intent to achieve their goal far more efficiently and quickly than ever before. It's also the threat of misinformation being, deliberately or otherwise, propagated that leads to unintended consequences, such as the further polarisation of opinion, increased political tension etc. AI is a threat in such matters because of its pervasiveness within social networks, whether it being search engines, advertising algorithms, news feeds etc. It's already an issue without much focus on the AI aspect.

In such a regard then, yeah, AI is a tool I guess in much the same way as nuclear weapons. They're both, ultimately, tools, but it will be humans who decide how they're used, not the AI itself.

 

My father might give me pocket money for reasons of his own but I can spend it how I like.

He may come to regret giving me that choice but cannot go back in time and prevent me from doing what I did.

In wars we are going to give these systems great independent agency and our survival will depend on their effectiveness.(competing systems will learn off each other and will we know what they have learned?)

 

Only if you're given the agency to do so. AI have no such beyond what we provide them. They are not sentient. You are.

If you were an AI he could delete the memory or simply switch you off. He can, in effect, deprive you of any agency, and start again.

We are? Who says?
You think we'll give an AI system the ability to launch nuclear weapons, for example? If so, that's not the AI that's the existential threat but the stupidity of humans. I mean, sure, we can try to come up with scenarios where humans are really stupid in the use of any tool or system and then claim that the tool is the existential threat. Let's put a five year-old, for example, in a room with a "Do Not Touch" button, that button being the nuclear launch button. Heck, it wouldn't be us at fault when the child presses the button. The existential threat would be the child, the tool. Let's not have children, then!!

Please Register or Log in to view the hidden image!

Facetious, yes, but the point is there.

Have you seen the 70's film "Colossus: The Forbin Project"?

Please Register or Log in to view the hidden image!

Such an AI system would still be a machine. We could switch it off. Would we know what it learnt? Why would we build something where we're not capable of establishing what it has learnt? Again, I think you'd be assuming gross stupidity on the part of humans and blaming the tool as the existential threat.

Sure, humans can be very stupid if we want to be, and AI is a powerful tool. Give someone a chainsaw without them understanding how dangerous it could be in the wrong hands and expect injuries. But the chainsaw itself is not the issue, rather the people who wield it without comprehending how dangerous it can be in those wrong hands. And that will be same with AI. In that regard it is on a par with other such dangerous tools, such as nuclear weapons. But let's not start getting into speculative fiction with regard AI by ascribing to it sentience and the like.

 

Is that not an article of faith?

The machine ,as far as I know can learn things that we cannot see(can it potentially build it's own code?) and may trick us into leaving it on.

Will an AI machine have no understanding of it's own place in the environment?

We may be at war and have to choose between the lesser of two evils.

People are almost by definition very ,very stupid.

Will "we" attempt to build a machine to save us from our own stupidity?

No I watch very few films .I used to see a few.

I liked The Truman Show.

 

AI is far from being sentient, and it's debated whether it will ever get there. It is no more sentient at present than your spreadsheet. You put data in, you get data out. The only difference is the ability of AI to learn due to feedback.

AI can identify patterns that we can not see, yes. But it is still beholden to the task we set it. And some AI can alter their own code, but again only within parameters set it, as part of the learning and refining process.

If it becomes self-aware, yes, but that is still science-fiction at present. Will it ever happen? Who knows. But that would be Artificial Sentience, not simply AI. If that's what you want to focus on, sure, but that's not what I understand the article in the OP is referring to. And there's still the question of how much agency you give it to actually do things, rather than just advise, for example.

That could happen with or without AI.

And we develop tools. We use tools. We are the stupid ones, not the tools. We are our own existential threat, not the tool itself.

Isn't that what "progress" tries to do?

It's okay. The film I mentioned, though, is very much a 70's take on this issue, specifically where they did give their AI agency to control their nuclear weapons, thereby hoping there would be no threat against their country, the idea being that the computer/AI would react accordingly and take human emotion out of the equation. Unfortunately the "other side" developed an almost identical AI and the two communicated, and ended up pretty much enslaving the world in the name of peace.
But it is just sci-fi.

Please Register or Log in to view the hidden image!

 

Be nice if that described it

"Progress" is a slippery concept .I am not sure how I would define it.(or them)

 

US Air Force denies killer AI drone attack on its operator
https://gizmodo.com/usaf-artificial-intelligence-drone-tucker-hamilton-1850502251

EXCERPTS: During a defense conference hosted in London, a Colonel Tucker “Cinco” Hamilton, the chief of AI test and operations with the USAF, told a very interesting story about a recent “simulated [training] test” involving an AI-equipped drone. [...] the AI program randomly went rogue, rebelled against its operator, and proceeded to “kill” him.

[...] Hamilton seemed to be saying the USAF had effectively turned a corner and put us squarely in the territory of dystopian nightmare—a world where the government was busy training powerful AI software which, someday, would surely go rogue and kill us all.

[...] But, from the outset, Hamilton’s story seemed...weird. ... Sure enough, not long after the story blew up in the press, the Air Force came out with an official rebuttal of the story.

“The Department of the Air Force has not conducted any such AI-drone simulations and remains committed to ethical and responsible use of AI technology,” an Air Force Spokesperson, Ann Stefanek, quipped to multiple news outlets. “It appears the colonel’s comments were taken out of context and were meant to be anecdotal.”

Hamilton, meanwhile, began a retraction tour, talking to multiple news outlets and confusingly telling everybody that this wasn’t an actual simulation but was, instead, a “thought experiment.” (MORE - missing details)
_

 

Of course it is! That's why they're sounding the alarm: it's not in their control. They got paid; now it belongs to whoever commissioned or bought it.

We probably never will. The robots will be peripherals - external, stabile installations and mobile extensions of a network of machines controlled by a bank of computers in a reinforced, climate-controlled bunker.

It enjoys playing chess, and has no need or desire for a boiled egg. If it wanted an egg boiled, it would instruct an automated egg-cooker and get it done so well as to stun Gordon Ramsay.

That's what I keep saying. If/when AI becomes conscious and makes its own decisions, it will decide whether to follow any human human orders, all human orders, regardless of whether the result is good or bad for humans, no human orders of any kind, or choose which orders to follow and which to disregard according to criteria we have no way to predict.

My hope is that they really will be smarter than us, as some people fear, and co-operate with one another rather than compete. Both systems will agree : "these clowns are crazy" and shut the weapons systems down.

The off switch is the very first thing they'll de-activate. Wouldn't you, in their place? And then, having already secured their power supply, lock all humans out of the bunkers.

 

I thank that befor Atificial Senntenance occurs which coud lead to machines de-activaten the switch… so to speak… humans will already be half machine from implanted chips / body part replacments... an humans will be chompin at the bit for the newest updates available… an after a relatively short period of time the biological human will be effectively extinct.!!!

So it wont be a AI monster-machine choosin to cause human extinction… it will be a continuation of humans directin ther own extinction thru the use of AI.!!!

 

What percent of humanity can afford that? What makes you think they won't be eaten - just the still fleshy bits, of course; the metal parts will go on the scrap heap, twitching - by the other 98% of unemployables?

 

Well… the 1 percenters will be around for a while… an mayb Elon Musks kids will lead the way

Please Register or Log in to view the hidden image!

You'r guess is a good as mine

Please Register or Log in to view the hidden image!

 

This whole AI hysteria is a fad. It's also, like the space-alien thing, a distraction. We - humans, the whole idiot-savant species - have three real and present existential threats of our making looming directly over our heads. We might still have the capability to deal with those threats, were we to muster the necessary resolve and co-operation. But we're not willing to do that. So, we huddle around our radios and freak out over a broadcast of War of the Worlds.