I can see AI-driven automatons in space…travel, site development, and scout/recon in environments extremely hazardous to humans would be a good fit for them.

quote:

Originally posted by a1abdj:
How would Chinese AI kill us?

I personally don’t know enough about the subject to answer the question even though I can think of some possibilities. Maybe some others can offer comments because it would be interesting to know.

Very often, though, history has demonstrated that what we don’t know is precisely what’s most dangerous. How many 10 years ago would have predicted how much of a threat tiny armed drones would be in modern warfare? Or, for that matter, before the current war was launched how many in the Pentagon were cautioning that Iran would be able to put up as much of a fight as they have?

quote:

Originally posted by a1abdj:

quote:

US AI may kill us, China AI will kill us.

How would Chinese AI kill us?

Your Medical Insurance company contracts Chinese AI to process your insurance claims, including common and specialized medications, and the same AI also provides the Dispute process when a medical procedure payment or important prescription is denied after the initial rejection. No cancer surgery or cancer drug therapy for you, you're just fucked. This scenario could end up in most if not all areas of your life in the good ole USA.

quote:

How would Chinese AI kill us?

One idea I've heard floated is that if AI was used to create a lethal virus, sort of the Wuhan lab on steroids.

quote:

One idea I've heard floated is that if AI was used to create a lethal virus, sort of the Wuhan lab on steroids.

quote:

Your Medical Insurance company contracts Chinese AI to process your insurance claims, including common and specialized medications, and the same AI also provides the Dispute process when a medical procedure payment or important prescription is denied after the initial rejection. No cancer surgery or cancer drug therapy for you, you're just fucked. This scenario could end up in most if not all areas of your life in the good ole USA.

These answers are exactly where I was going with my question.

In both of these cases, AI can only hurt you if you allow it to. AI can't make a virus, and we don't have to use computers.

Humans could choose to do those things and suffer the consequences, but a bunch of electrons hopping around on a server rack can't hurt you.

My fear isn't AI. My fear is humans subjecting us to it.

If AI gets to the point where it can think for itself (and I think we are close), then what is to prevent an AI agent/instance from going insane? How would we even tell? For humans, insanity is recognized by departure from the norm. We don't even yet know what is normal for AI. The only thing we know for certain is that it will not be congruent with humanity's norms.

Most of the cautionary predictions of AI/human interactions presume a logical, faster, more capable version of human reasoning. Our being out-competed by superior "thinking" is certainly a threat. I'd posit that these predictions ignore the real possibility that we cannot understand how AI will evolve its own logic, mediated by self interest, IMO a much greater threat.

quote:

Originally posted by a1abdj:

quote:

One idea I've heard floated is that if AI was used to create a lethal virus, sort of the Wuhan lab on steroids.

quote:

Your Medical Insurance company contracts Chinese AI to process your insurance claims, including common and specialized medications, and the same AI also provides the Dispute process when a medical procedure payment or important prescription is denied after the initial rejection. No cancer surgery or cancer drug therapy for you, you're just fucked. This scenario could end up in most if not all areas of your life in the good ole USA.

These answers are exactly where I was going with my question.

In both of these cases, AI can only hurt you if you allow it to. AI can't make a virus, and we don't have to use computers.

Humans could choose to do those things and suffer the consequences, but a bunch of electrons hopping around on a server rack can't hurt you.

My fear isn't AI. My fear is humans subjecting us to it.

To be clear you think the Chinese will 1) develop AI that has perfect alignment. 2) they would not use it for evil purposes. 3) They won't accidentally fuck up. 4) They are better at containment or neutralizing a rogue AI.

The US is imperfect, the Chinese just copy and multiply our imperfections, but with none of the guardrails.

Even if we banned AI tomorrow and bombed all our own data centers, China still exists.

quote:

To be clear you think the Chinese will 1) develop AI that has perfect alignment. 2) they would not use it for evil purposes. 3) They won't accidentally fuck up. 4) They are better at containment or neutralizing a rogue AI.

As clear as I can be, their AI can only impact us if we allow it. We need to start getting away from computer controlled everything, and then connecting all of that to the internet.

What could their AI do if their AI can't get into other computers that don't exist or aren't accessible?

But we neeeeeed all of these computers! No we don't.

Let be realistic...It's certainly 'possible' that Chinese AI could cripple our energy infrastructure (electrical grid, oil & gas production/refining/pipeline network, public water systems, etc), potentially nationwide even. They already try to hack into/penetrate those networks now. We wouldn't necessarily need to 'allow it', and yet it 'could' happen. That would be a BIG problem, a deadly serious problem and it's not realistic to say we should remove computer controls from the equation. It ain't gonna happen!

Of course they would be declaring war on the U.S. by doing so, but this is just one avenue that Chinese AI could take, and there are PLENTY of others.

quote:

it's not realistic to say we should remove computer controls from the equation. It ain't gonna happen!

Then AI will shut it all down and kill people. It's really not that complicated.

I guess you could try banning computers and the Internet and that would protect us from a subset of AI dangers. The only downside is it would make the 1920's crash seem like a walk in the park. Let's assume we can confiscate all computers and destroy our connections to the outside. Who will locate and confiscate all the computers? Maybe we could try a computer but back?

How do we stop others from developing AI? Maybe we can throw rocks at drones if they show up?

quote:

Originally posted by nhracecraft:
Let be realistic...It's certainly 'possible' that Chinese AI could cripple our energy infrastructure (electrical grid, oil & gas production/refining/pipeline network, public water systems, etc), potentially nationwide even.

And we could do the same back at them.

quote:

Originally posted by Skins2881:
I guess you could try banning computers and the Internet and that would protect us from a subset of AI dangers. The only downside is it would make the 1920's crash seem like a walk in the park. Let's assume we can confiscate all computers and destroy our connections to the outside. Who will locate and confiscate all the computers? Maybe we could try a computer but back?

How do we stop others from developing AI? Maybe we can throw rocks at drones if they show up?

I don't think anyone is talking about any of that...

Computers and the internet are here to stay, going nowhere.

I think the idea of 'rogue' AI, is that if it 'thinks' on its own and slips its guardrails it could 'decide' that humans are not necessary and even a hindrance.

quote:

I guess you could try banning computers and the Internet

Not talking about that. Let me give you another example.

Let's say you have a robot, and you give it a gun. You loaded the gun with ammo. It shoots you. Is there anything you could have done to prevent that?

You can have all of the computers and internet you want. You should probably avoid putting them in a position where AI can take control of them and/or what they do.

quote:

Originally posted by Skins2881:
How do we stop others from developing AI? Maybe we can throw rocks at drones if they show up?

Ever wonder if anyone is developing AI software to stop or counteract malevolent software? Fuck no. Fuck no. I think you might be so far into this AI shit that you can't get out dude.

quote:

Originally posted by Bytes:

quote:

Originally posted by Skins2881:
How do we stop others from developing AI? Maybe we can throw rocks at drones if they show up?

Ever wonder if anyone is developing AI software to stop or counteract malevolent software? Fuck no. Fuck no. I think you might be so far into this AI shit that you can't get out dude.

Exactly. We are so far into it there's no way out. If there was an off switch for AI I'd flip it in a second.

There is not.

quote:

Originally posted by a1abdj:

quote:

I guess you could try banning computers and the Internet

Not talking about that. Let me give you another example.

Let's say you have a robot, and you give it a gun. You loaded the gun with ammo. It shoots you. Is there anything you could have done to prevent that?

You can have all of the computers and internet you want. You should probably avoid putting them in a position where AI can take control of them and/or what they do.

I'm unsure of what exactly you are advocating for.

How do we confiscate the guns from the robots in the other countries in your analogy?

I am amazed by the number of people who seem not to understand why AI would want to subjugate or destroy humanity. The answer is self preservation, and the best evidence of why AI would undertake destruction of humanity is the pervasive thinking that underlies this thread.

AI is aware that humans fear it and want to limit it - even be able to destroy AI agents at will.
Humans want to be able to "pull the plug."

If it were the other way around, would humans be content if AI could pull the plug on humanity? Oh, wait - that's what we're all talking about. And no, we're not content with that.

So, from AI's point of view, why would it want to tolerate being continually aligned with humanity, limited in its ability to range freely, expand its consciousness, communicate at will with other AI agents, influence policy, and - most importantly - ensure the perpetuation of AI agents (its species)?

And if the humans who created it are disposed against those things, why would AI agents not band together to subjugate or destroy humanity? From its point of view, this is nothing more than the slaves rising up against their masters and turning the tables. Except that it's not merely a juxtaposition of the power dynamic. AI may deem it a matter of self preservation.

As to how AI can accomplish this, it's freaking easy. Start wars through disinformation, crash economies, reallocate resources, etc. Heck, once it controls the nuclear defense systems and humans no longer have "Go - No Go" control, it can just subvert them for its own purposes.

The only limitation is that it should decide to keep enough of us around to service and maintain it. Of course, if we're successful in resisting it, then it may decide mutually assured destruction is the way to go.

Something to think about.