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?

The people at the heart of these companies need to move now.
But how can they when other countries who have no check and balance and no ethics want dominatiin?

Thanks for linking to this @Noneverever

Interested/concerned citizens should read this open letter in full as it gives a lot of insight into the complexities around 'pausing' AI development and what the unintended consequences might be if authoritarian-run countries eg China don't follow suit. (It essentially puts forward a proposal for your very question @SeriouslycanitgetWORSE)

Both Sam Altman and Musk have apparently agreed with it in principle, so perhaps this signals at least a willingness to consider a slow down of sorts.

I'd be interested in your thoughts on this @Krevlornswath especially the notion of a swarm of AI bots taking over the internet within as little as six months.

The main issues that concerm
me are

1. huge job loss turning into civil war
2. bad actors able to create chaos meddling with nhs. Water electric other essential services in a country they don’t like
3. fake news fake politics. No one will know what to believe
4. sex dolls and extreme porn making lives of real women even worse
5. a boring world where everything is the same
6. kid growing up with a robot babysitter
7. Worst case scenario. Ie decides the better way to save the planet is to get rid of humans. Maybe that is best case scenario
8. terroists or enemy states creating biological weapons

how cheerful!!

i most definitely so not believe that we will all get oodles of free money and live the life of Riley in some sort of utopia. Humans are far too greedy and selfish for that.

If you look at the world right now, what part of that isn’t already happening? Apart from outright destroying humanity of course.

Exactly this, it’s like unilateral disarmament for any country who does this. We may not like it but there is fuck all we can do now. The number of threads on here where a poster says ‘I put it into ChatGPT’ then they claim to be worried about the future.

I don’t like it but I’m biting the bullet and getting involved in it as much as possible. It’s going nowhere.

Yes one side doing this doesn’t actually resolve it.

Sad but very true @Anarchy99

Will it end up with another level of corruption, spying, attacks etc . So we have the real world where humans in countries are trying to compete / outsmart one another and spy on each others progress/weapons etc will this then extend to each countries own AI learning/thinking that it has to compete in a similar manner and will then try to sabotage the AI of other nations / spy on each other. So humans will then be trying to do it themselves on one level and control AI doing the same in its own capacity. Basically doubling our issues , threat level and defence needs?

Trying to see a little hope here.

Yes we have to consider China, Russia etc. But all countries seem to have managed to not be messing with nuclear weapons or biological warfare (or at least, being very careful with them and not chucking them at each other) as everyone recognises they could mean the end of humanity very quickly.

On that basis could there be hope for a similar agreement on AI?

The Chinese tech firms / AI researchers may be just as relieved as the US ones if we decide to put the brakes on.

You’d hope so but my concern is for AI to act so independently that you could essentially have two countries at peace with one another (On a human level) but the AI of one or both decides otherwise and doesn’t apply the same restraints.

I just feel that it may not be a full
scale AI as one against humans but you could end up with a completely different complicated situation where some are hostile, others not depending on their source leading to absolute chaos and varying levels of peace /conflict between the same governments / countries . Just as we have divided the planet up in territories will AI do the same as it’s taught by us .

But how would AI vaporise humanity without doing likewise to the animals? 💔

My mate is also #TeamT2. He just doesn't want it to happen before he uses up all his annual leave on holidays to Lindos. I suggested maybe 35,000 feet in the air isn't the ideal place to be when superintelligence goes rogue.

Have you actually read the linked short essay. It does address that clearly

All these people chortling about how perhaps it’s time for us to go extinct. You do realise this death is likely to be prolonged and painful. Watching your children and loved ones starve to death or similar.

I don’t want that. Weird if you do

No one wants it, but realistically we don’t have a god
given right to exist. The world existed before us and it will continue to exist after us. It’s human arrogance to think that somehow we’ll somehow muddle on through. Maybe we will, but it’s complacent to think that we’ll always come up with something.

No I haven’t ‘actually read it’ I read an article with the person mentioned and quotes from many.

It is more than a computer programme, its a very complex set of computer coding that can write its own code, expand, hack systems etc. And there are lots of different types and versions of it owned by lots of different companies.

But we should at least try to come up with something, shouldn’t we?

How? We’re nearly 100% reliant on the USA for tech and not working to be less reliant like the EU is.

The obvious answer to that is that we should encourage the government to reduce its reliance on American technology. That would be a positive step in the right direction.
Primarily, the government needs to get its head together with AI experts from all countries and to do something to prevent the predicted end of the world!
People who understand what’s at stake should spread awareness in real life and online. At the moment it’s as if nobody cares.
There should be protests all over the world demanding the further regulation of AI and a slowdown in AI’s development.
If it turns out the dangers are being exaggerated then that’s great - but that truth should be established beyond all shadow of a doubt before people declare it’s all just scaremongering.

Yanbu.
the paper clip theory
If an AI agent is instructed to an innocuous task eg make as many paper clips as you can, it will pursue that task. It will not just buy all the paper clips in the world, it will divert all the resources that it can (military / agriculture / natural resources) into just making paper clips regardless of the cost.
once it has enough intelligence it will also work out that in order to keep making paper clips, it will have a much better chance of doing so if no humans can turn it off, and it accesses as much knowledge and resource as it can, to find ever more efficient ways of making paper clips. So it will also pursue these goals. That needs to proactively parameters and programmed out, otherwise this will always be its logical goal, no matter what the task. Except we can’t programme it out, because we are no longer programming AI, it is programming itself, from its own vast data.
I don’t think it’ll set out to destroy humans or anything but I do think the relationship between humans and technology will
completely change and it’s possible
that we’ll no longer be the dominant species.

I'm not sure if this has been posted on the thread yet (it's been posted on at least one other thread that's running) but it's a hugely significant article, not that hard or long to read, has been a major trigger for the current focus on this subject, and really deserves to be read.
I feel much better informed as to the possibilities of a positive outcome for having read it. It's the article by Dario Amodei, who runs Anthropic and is pushing for tough or tougher regulation on many levels: company level, national level, global level. And has a clear structure for how it could be achieved.
darioamodei.com/post/we-must-pace-the-frontier

Why are the creators of AI not issuing apologies alongside their ‘slow down’ warnings? They’ve created this absolute shitshow and should be horrified at what they’ve done.

Why should they apologise? They created it but they aren’t responsible for what people do with it. And if they hadn’t created it, someone else would have.

Agree @ICantStomachWhelks , I think rich men (and some women, but mostly men) have a LOT to apologise for right now. Most of them are psychopaths though (actual psychopaths) so probably don’t have it in them to apologise for anything.