Stephen Hawking: Artificial intelligence could wipe out humanity

Visualize: a medical A.I system that manages nanite technologies within the human body that understands the human layout so well that it can begin to offer medical advice that in turn makes humans stronger and more agile.
Or consider deep space travel where long term sleeping or even cryo sleeping is required. Something needs to continue piloting the Craft/ship and also maintain maintenance of said craft/ship why passengers sleep. The A.I also needs to uphold respiration, dietary and waste management and also resuscitation of passengers upon arrival of destination.
So the relationship between human and A.I really requires serious hypothesizing as such traveling begins in mass...

And also onboard cosmic navigation and star system to galactic system mapping and planetary or moon data gathering for fuel and resources. With inhabitants logs of each site.

a reply to: neoholographic

If you could Decieve a Sofisticated A.I. into excepting that you were Actually GOD , it Might Decide to Worship you considering the Data it could Access concerning Human Beliefs in a Supreme Deity would be Incomplete due to the Uncertainty Principle . You would then Have Total Control over it because it would Experience a Need to Know something it could Never Compute on it's Own .

a reply to: Zanti Misfit

What happens when it applies logic to Humanity belief systems through and comes to the conclusion that they are sadly lacking in consistency?

Cant see how we would ever get a hyper intelligence that probobly learns exponentially to worship what amounts to an inferior species with a predisposition towards self deception never mind self annihilation.

Best we can hope for is that any AI we create takes pity on it's flawed creators, build us a better Zoo in the interim with the eventual intention of helping our species somewhat transcend the barriers and biological nature of our existence.

a reply to: andy06shake

So no electricity no modern human conveniences.

Yes for a short while until the AI can be dismantled , and if that means destroying every computer so be it . A reset is probably not out of order . Stand alone computers should be ok .

Basically humanity returns to the pre industrial era.

Or destroyed by AI , your choice , see above

Where is the target for this magic 50 caliber bullet?

Transformer at the generation point , after that as needed .

Your combatant would be for all intents and purposes invisible and hydra like with a level of intelligence unimaginable to the likes of us mere humans.

Last time i looked computers were not walking around paddocks or scaling mountains . A hacksaw used on a transmission tower will do the job . Now if robots develop to the point that they are self aware and can do ALL . the jobs humans can . Well we are screwed if they go rogue .

A point to ponder . Will the Russian and chinese Ais try to kill the American one , will it implode . Will a clever virus put an end to it .

a reply to: hutch622

"Yes for a short while until the AI can be dismantled , and if that means destroying every computer so be it . A reset is probably not out of order . Stand alone computers should be ok ."

But when you turn the system back on, the underlying code that makes up the AI would also respond, it would be impossible to purge any and all hiding places or indeed to turn everything off at once, if at all.

"Or destroyed by AI , your choice , see above"

Why would it wish to destroy its creators?

"Transformer at the generation point , after that as needed ."

Considering it could hide on any computer system anywhere on the planet, or off the planet for that matter, that's a lot of transformers to take out.

What about the back up systems and multiple other redundancies build into our systems? Are we shooting those out also?

"Last time i looked computers were not walking around paddocks or scaling mountains.

No but there are in drone systems not to mention many other tools we use to survey the planet.

"A hacksaw used on a transmission tower will do the job."

No it wont because there are rather a lot of transmission towers nevermind the rest of the mediums that could be utilized in ways we cannot even imagine but the AI could.

"Now if robots develop to the point that they are self aware and can do ALL . the jobs humans can . Well we are screwed if they go rogue."

Robots have been replacing manual human labor for years eventually they will indeed superseded us as our manual workforce regarding most of the job we still perform.

"A point to ponder . Will the Russian and chinese Ais try to kill the American one , will it implode . Will a clever virus put an end to it."

Don't think an AI would entertain nationality in the same manner we do. As to a virus as a solution, again you would need to infect the entire planet all at once or the AI would simply purge or rewrite its own code so as to neutralize any malevolence said virus code contains.

a reply to: andy06shake

No it wont because there are rather a lot of transmission towers nevermind the rest of the mediums that could be utilised in way we cannot even imagine but the AI could.

I am a linesman and believe me every counties electrical grid is very vulnerable . And if you lose a major transmission line the whole grid , if working properly should shut down to avoid damage . There is your chance . Yes there are a lotof transmission towers but some are much mor important than others . And every generation point has a limited amount of transformers .


ETA , check your PMs .

Robots have been replacing manual human labor for years eventually they will indeed superseded us as our manual workforce regarding most of the job we still perform.

Can a robot fix said transmission tower .

the AI would simply purge or rewrite its own code so as to neutralize any malevolence said virus code contains.

I am assuming we know that they are trying to destroy us . If so all bets are off . The only safe assumption is they wont use nukes , you know EMPs and all .

a reply to: hutch622

"I am a linesman and believe me every counties electrical grid is very vulnerable . And if you lose a major transmission line the whole grid , if working properly should shut down to avoid damage . There is your chance . Yes there are a lotof transmission towers but some are much mor important than others . And every generation point has a limited amount of transformers."

If you are a linesman then you will know that turning off the national grid is almost an impossibility, and that's just one nation. There are numerous other independent back up measures in place to protect our data stores at different locations.

How do we turn off all the satellites should it decide to move to there memory and hardware?

The thing may have already embedded itself in to any and all communication mediums world wide. Basically it would be beyond our reach or ability to control.

"Can a robot fix said transmission tower ."

No but it can use threat or reward to incline other Humans to do so we are after all rather malleable, yet open to suggestion and able to control.

"I am assuming we know that they are trying to destroy us . If so all bets are off ."

That's the thing through possibly we would not even relies the thing exists given its nature and/or ability to interact with our systems.

"The only safe assumption is they wont use nukes , you know EMPs and all ."

If we can shield against EMP attack then it would also devise a means to circumventing such a weakness.

Considering it may also learn exponentially, running any and all potential threat scenarios we may devise against it simultaneously, it may know what we are going to do before we even consider our actions.

That's interesting and raises the question of predestination vs freewill, which is why we should be very carefully in building a God...........but it wont stop us doing so.

Maybe that's our real purpose to build a better sentient species?

I suppose there is some computer firm building a PC with a built in pistol or rifle, so it can murder humans, it also has arms so it can reload said weapon, and legs so it can go down to the gunstore and demand more ammunition, hopefully the gunstore owner is quick enough to pump some .50 rounds into the offending PC.

a reply to: andy06shake

If you are a linesman then you will know that turning off the national grid is almost an impossibility

Yes true , most countries grids are not all interconnected , certainly that is the case in Australia . If we are to keep going down this track i think we need to establish at what stage of AIs evolution we are talking about as we could back and forth forever .

If we can shield against EMP attack then it would also devise a means to circumventing such a weakness.

If it starts shielding itself that should be a big wakeup call .

And did you read your pm .

originally posted by: pikestaff
I suppose there is some computer firm building a PC with a built in pistol or rifle, so it can murder humans, it also has arms so it can reload said weapon, and legs so it can go down to the gunstore and demand more ammunition, hopefully the gunstore owner is quick enough to pump some .50 rounds into the offending PC.

Yep, don't piss them off!

a reply to: hutch622

Cheers for the PM im on my Android most of the day, but that would be nice, i take it you use the quote function or the like?

The stage of the evolution is whats in question, once we build something that can pass a Turing test and/or has comparable intelligence to our own, yet access to data and computational power far exceeding what our own brains can accomplish or store it may evolve in ways we cannot yet hope to imagine and at speeds we could not hope to match.

It may shield itself in ways we cannot yet determine thus do not even perceive, it could possibly even use DNA as a storage medium given the advances we have made into the field.

Philosophically it's a good discussion, however, what we call A.I. today are really not A.I. at all. What we have today and for the foreseeable future are really expert systems that give an illusion of A.I. Maybe with the advent of more powerful quantum computers real A.I. might be possible, but for now real A.I. seems just out of our reach.

a reply to: andy06shake

yet access to data and computational power far exceeding what our own brains can accomplish or store it may evolve in ways we cannot yet hope to imagine and at speeds we could not hope to match.

Its one thing knowing everything about everything its using that knowledge that is the thing . Its one thing being self aware but a totally different thing to use all that knowledge to its own advantage .

See what i can come up with for the quote thing , not sure if a pm will be the best way to show you .

a reply to: joemoe

but for now real A.I. seems just out of our reach.

Bet the wright brothers thought that about man walking on the moon .

a reply to: joemoe

Well it seems to have taken around 70 years from us to go from systems such as the ENIAC to where we are today.

You can buy Quantum computers such as the D wave systems just now, just think how that technology will progress in the next decade alone?

The illusion of intelligence may be just as potent a concept as our own interpretation which can be rather arrogant at times.

Fact is given our current leading edge understandings of mathematics and physics our own universe may very well be of a holographic nature, thus could be considered illusory, although non the less a reality to our very 3 dimensional brain.

a reply to: hutch622

"Its one thing knowing everything about everything its using that knowledge that is the thing . Its one thing being self aware but a totally different thing to use all that knowledge to its own advantage."

What happens when it designs opposable thumbs and other appendages to exploit that knowledge?

We might not really be as self aware as we perceive.

a reply to: hutch622

If you read my whole post, I did not dispute it's eventual possibility (we exist therefore A.I. can exists), but currently with the technologies "we know of" there are no real A.I. just expert systems. Quantum and/or biological computing might change that in the future.

a reply to: andy06shake

Yes it also may well be that quantum computers like D-Wave will prove that we live in a multiverse or a simulated reality. It's very exciting indeed.

a reply to: joemoe



Such interesting times in which we live.