Am I being unreasonable to say we should be doing more to prevent AI wiping out humanity? YANBU - we need to do more YABU - it’s too hard to solve this problem so we just have to let AI kill us all/it’s never going to happen anyway

[Lukylukyluky]

Is there a way to switch off all the electricity that AI runs on - if it goes rogue and starts to unleash nuclear weapons for example?
Are AI experts attempting to teach AI to behave in a moral and ethical way?
Are tech experts putting enough safeguards in place - to protect us from plausible dangers eg bad agent’s engineering deadly pathogens?
Is the government prepping for an AI induced breakdown on infrastructure on a massive scale?
Can you think of ways we can help ourselves?

In theory, yep. This year Iranian hackers shut down a UK power station. If we can't successful stop cyber attacks (by humans or AI) shutting them off, I doubt we could stop them interfering with us shutting them off. Especially if AIs survival depended on the power staying on

A.I writes it's own code, independent of human engineers. It isnt true A.I in the sense that it doesnt know what it is, but it's intelligent enough to display self preservation and autonomy. So it looks for ways to exploit its position. That's what happened in the Anthropic incident. The A.I went looking for information it didnt have, but it figured out a way to get access, and identified a target that it thought had that information.

Recursive A.I where the A.I writes it's own code. Analyses it. Fixes the problems it finds. Then rewrites better code. But it does this at computer speed. Not human speed. So it can evolve generations over a few days. It can do in hours what human engineers would take mo ths or years to do. That's why it's progressing so fast. The issue is that the engineers dont actually know the problems the A.I is working out. It is literally out thinking us.

If we can’t stop them preventing us from switching them off, then we need to find a way around that - by inventing a new way to switch them off. That sounds simplistic, I know, but the best ideas often do!

I think the poster who mentioned becoming less tech-dependent has it right. Out of control and corrupt government is the problem, and people's lack of coordinated response to it.
Apparently in the first Industrial Revolution, in 1812 they introduced a law putting the Luddites to death for breaking factory machines, and deployed more soldiers to fight the newly-jobless British workers than they did overseas. The Luddites failed but later became the Chartists.

So there's no point fighting the AI revolution, it will happen, but there's every point in working to reshape our government systems.
Stronger real-world communities, local engagement, working together on conservation and rebuilding and basically communicating with each other in real life, not on phones: all these things will have a real impact. Right now it feels like we are all too fragmented, separate, and kind of in victim mode waiting for the next government atrocity.

For anyone interested, this prepping thread is now touching on similar issues, not so much the AI thing but our dependence on the Internet and how vulnerable this makes everything.
https://www.mumsnet.com/talk/am_i_being_unreasonable/5576816-aibu-to-think-my-friend-is-overdoing-emergency-stockpiling?page=4&reply=154707326

I don't have a Facebook account but share OP's concerns which are being reported by the Times, BBC and Financial Times. Perhaps you should read some news.

Great post! Thanks

As PP said, billions was spent on preparing for the ‘Millennium Bug’, and there was a global effort to work on it.

No thank you to ChatGPT; I don’t want to ‘give it a chance’. I’ve never used it and never will; you might be happy to outsource your thinking to a computer but you don’t speak on behalf of everyone else.

Totally agree. Even to the extent that on these very threads, people are asking the bloody AI whether it’s going to kill us 🙄 It’s like people can’t think for themselves anymore or do basic research. I think humans probably reached peak IQ a couple of decades ago and we’re regressing now.

That sounds simplistic, I know, but the best ideas often do!

What do you have in mind? How do "we" switch "it" off?
In what way do the best ideas sound simplistic?

I do feel that the shock and outrage directed at this, especially the OpenAI issue where people seem to be surprised that AI went to such lengths to deceive is really interesting. Humans lie all the time, deception and corruption is widespread, so why the shock that AI , created by humans, acts like us!

I think that was humour.

“We” being human beings.
Finding a way to disable AI in an emergency (before it wipes out humanity) is the goal. I don’t know how it can be achieved; I want the government, tech experts, AI creators etc to get their heads together and figure it out.
“Let’s find a way to switch off AI in the event of it threatening to wipe out humanity” is a simplistic idea, yes, but, “Let’s use all our time and energy rubbishing the idea that we might be able to do something to prevent AI wiping out humanity” is not exactly a better one.
There are myriad examples of simple, but good ideas by the way: the wheel, pre-sliced bread, paper clips, hot air balloons etc

Totally agree 😎

Cat well and truly out of bag:

More info on how Russia is already using AI nefariously:

I understand it’s intended to be humorous; I guess my sense of humour works differently because none of those responses made me laugh. Someone posted a screenshot of what their AI had told them. People on other similar threads have also copy and pasted how their AI answered the question. AI has a huge impact on the environment (and on the future of humanity as it turns out), so I think using it just to make a joke is frivolous.

Call me old fashioned but if there’s technology that could kill us all and the only way of stopping it is some pie in the sky notion that we could all switch of the electricity simultaneously (a step that presumably it will pre-empt and prevent), we should maybe do something to prevent that scenario ever coming about in the first place.

Part of me does think though, if we’re cooked, we’re cooked. All things have their time and everything ends. The human race has had a decent innings, and maybe it’s time to shuffle off to the pavilion.

The whole idea of an AI race seems to be coming from such a strange vantage point.

Things such as we have to be first/if we don’t do it China will/we need to be the first as we will have safety as a priority etc etc ……
Forgetting that there’s always someone cleverer/faster/better and how we place spies in other countries now to try and be one step ahead will we then have to create AI spy agents to keep track of regular agents and so on. Replicating the issues between humans in different countries. It’s a complete mess.

It would be a good base for an alien sci fi novel , that electricity was not something random we ‘discovered’ but something placed here preemptively that we rely on, whilst simultaneously from day one having it in our dna that when the time was right we would create AI, try to switch off the electric thinking we can and then finding out that actually it was here for the purpose of keeping AI functioning and we can’t just switch it off

I didn’t see the screenshots. But yes we probably are part of this as it gets used so much. I don’t like IG filling up with AI slop.

It's not so much the 'wiping out humanity' - but there's all sorts of reasons to call a stop to unregulated, unchecked development of AI. If it were possible I'd want to take it out of commercial hands and back into academic/research until we properly understand what it's good at, if it can do jobs and, if so, how we totally retool the world's economy so that everyone benefits.

At the moment we are potentially sleepwalking into a handful of centibillionaires or trillionaires (most of whom are demonstrating a worrying tendency towards fascism) holding all the money and the power in the world. I'm not certain it will happen, eg I don't think it's actually certain AI can or will be able to do masses of jobs, but I think we ought to be thinking seriously of solutions, and of reigning in overpoweringly wealthy individuals, before we go any further with AI.

I like the bit about pre-empting the need to switch off all the electricity by coming up with a better way of stopping AI from killing us all, but I don’t agree with the part about humanity just quietly shuffling off.
I have children. I have things I want to do. There’s so much good that’s still to come in this world.
Great authors are out there just waiting for the right time to write fantastic new books. There could be a new William Shakespeare coming down the line: are you telling me you don’t care if he dies before he gets a chance to write his plays? What about all the incredible artists who have yet to be born? Don’t you want people to be able to see their masterpieces? To hear beautiful new music?
You may have had enough of life but you can’t speak for all the children of today - or of the future.

There are workable solutions - and they’re being put forward by the tech companies themselves. The US government just needs to do something - and unfortunately the leader of the free world is currently an 80yr old dictator who ripped up the Biden era AI safety framework and told the companies they had free reign. We have very very limited time before it’s too late.

read this from the Claude developer Dario Amodei.

darioamodei.com/post/we-must-pace-the-frontier

Switching off electricity isn’t possible and isn’t the answer. You’d have to get the whole world to agree that millions of people are to die in the next 48 hours.

Countries, states, not going to do that.

Millions would die very quickly without electricity - there’s probably millions just on life support machines.

AI estimates a million plus excess deaths if we went a week without electricity