I, Robot , I Soldier.

[Scottieoutdoors]

Had a thought, apologies if it's already been covered... we all seem to be talking about AI as if it's a human vs AI situation... we are all quite divided across the globe...culture clashes etc...

Could AI be coded to learn what is best not for the human race, but for individual countries or beliefs or cultures? Can we control the parameters in which it learns... could we end up with WW3 being started because one nations AI is deemed a threat to anothers? That to me seems more spooky/concerning than just AI vs Humans...

 

[Broch]

Highly probable

 

[TeeDee]

Had a thought, apologies if it's already been covered... we all seem to be talking about AI as if it's a human vs AI situation... we are all quite divided across the globe...culture clashes etc...

Could AI be coded to learn what is best not for the human race, but for individual countries or beliefs or cultures? Can we control the parameters in which it learns... could we end up with WW3 being started because one nations AI is deemed a threat to anothers? That to me seems more spooky/concerning than just AI vs Humans...

Or potentially a 'cleverer' A.I bypasses another A.I security protocol and hijacks it and its assets.

 

[Paul_B]

Or humans are clever enough to ensure AI doesn't destroy the planet. Perhaps by doing it themselves before AI gets a chance!

Sorry! Couldn't resist.

 

[Paul_B]

Seriously though, do you guys really think AI is going to be a big problem to the human species? Was other tech with dangerous potential not once thought of along similar lines but now isn't and hasn't been a much of an issue as once made out to be?

I sometimes think new and powerful technological steps gets greeted with fear, loathing and scepticism. Then cycles round to acceptance and appreciation even.

 

[TeeDee]

Personally I think the big question is if the machine is acting on your countries interests would you , as a citizen of that country be at peace with your conscience if the termination of another human being wasn't undertaken by another human??

I know that maybe a slightly weird way of looking and considering it but its what concerns me the most I think.

 

[TeeDee]

Seriously though, do you guys really think AI is going to be a big problem to the human species? Was other tech with dangerous potential not once thought of along similar lines but now isn't and hasn't been a much of an issue as once made out to be?

I sometimes think new and powerful technological steps gets greeted with fear, loathing and scepticism. Then cycles round to acceptance and appreciation even.

I don't think A.I falls into the typical chicken little type scenarios that have blown through our civilization to as you say be seen with concern before being mastered.

I genuinely think its going to be a monumental watershed moment - and the thing is alot of the visionary type people that have been working on it , more au fait with it capabilities and potential concerns are the one ringing the alarm bells as we seem to get closer to an edge of the apex type roller coaster moment.

I'm reminded of the following adage.

"Fire is a good servant but a bad master."

 

[Paul_B]

There's a lot of the visionary types involved in it who do not hold that view too. It's still not accepted fact that AI is the big risk or potential risk that media like to promulgate to sell their brand. I bet too that the proponents of this fear from among your visionaries didn't exactly say things as reported too. Any scientist worth listening to won't use such definites like that I reckon.

 

[Paul_B]

PS I think there was the same attitude to nuclear at one time. Where are we with that destroying the world?

 

[TeeDee]

Seriously though, do you guys really think AI is going to be a big problem to the human species? Was other tech with dangerous potential not once thought of along similar lines but now isn't and hasn't been a much of an issue as once made out to be?

I sometimes think new and powerful technological steps gets greeted with fear, loathing and scepticism. Then cycles round to acceptance and appreciation even.

There's a lot of the visionary types involved in it who do not hold that view too. It's still not accepted fact that AI is the big risk or potential risk that media like to promulgate to sell their brand. I bet too that the proponents of this fear from among your visionaries didn't exactly say things as reported too. Any scientist worth listening to won't use such definites like that I reckon.

You asked a question. I answered it.
Everything else is conjecture.

 

[Scottieoutdoors]

@Paul_B
The difference being with nuclear is that it requires a human for it to be used (unless malfunctions count...)

This is the talk of a computer or algorithm deciding to use it.

@TeeDee, I suppose I would liken AI to the introduction of a man made disease or a virus that was created with the best intentions, yet wreaks havoc when it's truly loose... or liken it to the possibility of that happening...

 

[Paul_B]

An algorithm written by a human. You think we'd create something knowingly with the potential to launch nuclear armageddon or other? AI is currently being used for data handling and analysis but not control uses.

Chernobyl, 3 mile Island, numerous leaks in UK stations, that pit in the far north of Scotland that last I heard was a big unknown wrt nuclear waste. Loads of dangerous situations and we're on the brink of another with the world's largest nuclear power station in a war zone and probably mined. That's real risks, real and known. AI is all conjecture and basically relies on clever ppl absolutely messing up for it to become an issue. It sounds like you're all assuming that they'll take that idiot's step and give AI the security codes and launch codes. It's bad enough the ppl we have with those, I don't think they'll give them to a homicidal AI program.

 

[Laurentius]

It has to be appreciated that given the law of averages, half the population has an IQ under 100, however I would estimate the current IQ of AI so far as I have seen it as very much less than that. I think people misunderstand what AI actually is and how it works. The basic rule applies regardless, garbage in, garbage out.

 

[TeeDee]

It has to be appreciated that given the law of averages, half the population has an IQ under 100, however I would estimate the current IQ of AI so far as I have seen it as very much less than that. I think people misunderstand what AI actually is and how it works. The basic rule applies regardless, garbage in, garbage out.

Not really - in the same way that computer power has expediently jumped in the last 40 years A.I will be able to accelerate its learning trajectory at a similar growth curve.

Yes , A.I maybe seen to be rudimentary today but remember since Mobile phones were once basic and simple look how quickly they have developed into cutting edge technology. Its said that a modern phone today has more computer power than the NASA control centre that helped the first Moon landings.

Tech evolves , quickly! A.I will no doubt increase at a similar rate of speed once unfettered.

Please feel free to explain what your definition or understanding of A.I is then? " I think people misunderstand what AI actually is and how it works."

 

[Scottieoutdoors]

@Paul_B
I sort of feel that you have a reliance on the human algorithms as an impassible barrier for AI...

I think people need to consider AI as a living creature... take a doggo, we get dogs from pups, we train them and mold them to the companion we hope them to be, but they are still their own creature, they can decide if they so wish to go completely off piste based on their free mind.

Now AI in its current form is, I'd suggest, like an extremely well trained doggo, it does the deeds it is asked to do, it figures out what people want it to do so on so forth, but it's designed to learn, and designed to do things at great speed...
A pass code with only 2 numbers has what, 100 possible combinations? How long would that take a human vs a computer to figure those out? 3 digits, 4...12?

I hope human coding can be watertight, but as already mentioned, we're pretty useless with nuclear containment and safety, only seemingly learning after the fact... can we afford that with AI?

Ofcourse AI has many great uses too, so it'd be handy not to lose altogether.

 

[Broch]

An algorithm written by a human. You think we'd create something knowingly with the potential to launch nuclear armageddon or other? AI is currently being used for data handling and analysis but not control uses.

That dismisses some of the more advanced adaptive algorithm AI that is being used (and has for some time) - AI that changes and adapts its algorithms, not just the data, depending on conditions; effectively writing it's own software.

AI is very much being used for control! I used AI type control systems (albeit without the data access and the computing power of today) in the late 80's!

However, the real dangers may not be an Armageddon type situation - it will be the perversion of truth and the total loss of privacy; it will be the automation of personal level decision making about access without anyone to appeal to; it will be the total loss of jobs in some sectors - already telephone support offices are being closed.

Exciting times - I love it

 

[Broch]

And then ....

More than 1,300 experts call AI a force for good

An open letter organised by the UK professional body for IT says AI is not a threat to humanity.

www.bbc.co.uk

 

[Glass-Wood-Steel]

Nobody can say for sure whether AI is a threat to us or not. It is here now and we will find out for better or worse. The more it develops the less chance we have to understand how something alien actually thinks. It is non human so our experience and way of judging things by our own standards kind of drops away. Many of Mankinds creations have a sort of Frankenstein factor to them. I do not mean all are monsters or monstrous, I mean there are flaws and twists to them however well the intentions used to invent them. This does not even mean they all go bad either, it it is based on the idea that once something is created for the first time, only then do you discover the consequences and once you give autonomous life to something it will be, well, autonomous.
I do not think it is wise to invent more and more things that either pamper us, remove challenges or where a decreasing amount of people even understand or know how to maintain.

 

[Glass-Wood-Steel]

There was an interesting idea that I read once kind of on the theme I was on in the previous post.

AI and the paperclip problem

Philosophers have speculated that an AI tasked with a task such as creating paperclips might cause an apocalypse by learning to divert ever-increasing resources to the task, and then learning how to resist our attempts to turn it off. But this column argues that, to do this, the paperclip-making...

cepr.org

This is a version of it, obviously a larger than life story but it illustrates the idea of not understanding the full ramifications before starting something.

 

[Kav]

I’m watching the southern USA from California to Georgia experience a killer heatwave. The northern part is coughing from Canadian Wildfire smoke.
CLIMATE CHANGE is accelerating faster than the best predictions a few years ago.
refugee crisis, wars, mass starvation , disappearing species and entire ecosystems , the spread of diseases, King Charles and Harry, Elon Musk and peers trying to run the word, gun violence in my country, trump won’t go away, but good people do...,
Uh, why are we worried about AI
Again?