'Killer robots' with AI must be banned, urge Stephen Hawking

[Emu]

Really we already know it's not going to be a problem, because if it were, killer robots with AI from the future would have travelled back in time to kill Stephen Hawking's mum.

[NineLine]

Really we already know it's not going to be a problem, because if it were, killer robots with AI from the future would have travelled back in time to kill Stephen Hawking's mum.

 

Well depending one what version of time travel you believe in, maybe they did and it just doesnt effect our timeline ;)

 

[Ryan Legend]

An AI would be able to think and create it's own ideas based of the data that it receives. An AI would be able to "program" itself and create dynamic algorithms, so it can "learn".

 

This AI isn't perfect, you can't learn something without making mistakes first. This is what practice is for. This machine is not so different from a humen being. You receive data from the world arround you all the time, based of that data you will make decissions. How you make that decission doesn't matter, you're algorithms are dynamic and they change all the time. An example of this is growing up. You're not the same person you were as when you were at the age of 8 or 12. This is because you learn, but to learn you have to make mistakes.

 

Now imagene a computer handeling a weapon, this computer is NOT intelegent. It is in no way like the AI that I described before. In fact, this piece of software is the exact opposite of the other one. It receives data and with that data it only has to fill in the blanks. (A fixed algorithm.) It's simple and thus there are no mistakes. It doesn't have to learn, or reprogram itself, it doesn't have self-awerness or emotions. It can make simple dececions autonomously based entierly of pure logic and mathematics. The dececion it will make will always be the best one. Because it is programmed to do that.

 

I'd rather have this computer operate a weapon than a very complex humen being, basing its dececions mostly on emotions, that has to make mistakes to improve itself and gets corrupted when he gets to much power. (god complex).

 

Its in our inner nature to go to war, machines don't have that.

 

just my two cents and personal cake to the party :)

Edited August 4, 2015 by Ryan Legend

[Emu]

Well depending one what version of time travel you believe in, maybe they did and it just doesnt effect our timeline ;)

 

So, in one version of the future, Skynet won but lost in two?

[104th_Maverick]

I agree that AI is pretty scary but what is straight up terrifying for me is the potential for these things without AI.

 

A cheap drone with GPS programmed information could do a lot of damage possibly even undetected, the potential for abuse of systems like these should not be underestimated.

 

These things will pretty much have the ability to enter any airspace on the planet apart from areas with highly sensitive perimeter systems but even then, a swarm of these could do much more damage to a classified installation than a regular external aircraft threat could do, especially since any aircraft would have to pass through several layers of defence.

 

Possibly not on the horizon at the moment but eventually threats like this will develop and someone or some government will undoubtedly try to abuse them. Talk about a False Flag... imagine a drone that can go anywhere in the world at low level silently in an urban environment, employ a weapon and leave without anyone ever knowing it was there..... scary!

[Exorcet]

On the flip side AI drones allows people to send swarms of never tiring defenders into the same airspace to keep watch of every single point of vulnerability.

 

Small micro drones are also going to be limited to small payloads, ranges, and speeds. It's an arms race like any other where threats emerge and counters emerge to fight the threats.

 

The problem with banning them is that it allows anyone who doesn't want to follow the bans to get a huge lead over everyone else.

[zaelu]

AI is no AI if it is still pre-controlled. is just a very smart weapon and the guilty is the "commander".

 

Apropos, if a tyrant would have a 100% robot army will be the easiest to defeat as it would be totally disconnected from his human enemy. Is the tyrants with a clique that were hard to defeat.

 

The single thing people need to understand is that AI as a.. robotic intelligence... meaning different from a "Cylon-like" one does not compete with human species on resources. Resources competing is the sole source of violence in this world.

 

Personally I appreciate the great minds that sign that paper o whatever it is but I think they are very very wrong in doing this and encouraging such fear mongering trend.

Edited August 4, 2015 by zaelu

[tovivan]

I'm afraid that no matter who protests and no matter who signs letters or how many millions of letters are written, the military will do what it wants regardless, even if it means covertly and even if it means covering the costs by claiming they needed thousands of new toilet seat covers in their bases and each one cost $5000 or $10000.

 

@zaelu: the problem here is that robots and AI's have no compassion. There is still a chance that an indoctrinated soldier will have a moment of weakness or compassion and let someone live or not fire a rocket or drop a BDU, but there is no such chance with robots. They will kill everyone their programs allow them, man, woman, child, seniors, mentally/physically challenged, puppies, kittens, maybe the entire world if they're not stopped. And when an AI becomes sentient? What if it decides the world would be better off without us destroying it?

Edited August 5, 2015 by tovivan

[zaelu]

As one said above... in the long run I think is better to have robots with better and better programs and also with less and less usage (because robots killing each other are a futile effort in the end) that to be at the mercy of "compassion" of a brain washed soldier activated by hate and sheer stupidity. You have to think in numbers although sounds bad. If autonomous (AI) drones would kill 100 children the whole earth (enemies and friends alike) would protest, but if a human party does that... many will find some political, sexual, religious excuse for it.

 

As for the last part. As I said. AI has no competition with humans on resources hence once they become sentient (as we like to be proud of ourselves) they will have no reason to use violence. I am thinking that even in defense there might not be necessary.

[tovivan]

because robots killing each other are a futile effort in the end)

 

Ever heard of the ancient PC game Wargasm? :D :joystick:

 

that to be at the mercy of "compassion" of a brain washed soldier activated by hate and sheer stupidity.

 

True story. One specific example that comes to mind is a friend of mine. He's a retired British Army Sergeant, a Cambridge educated man and I would even tentatively call him an intellectual.

 

But even with all that, the indoctrination he was infused with so many decades ago still holds completely firmly and he has never and will never shake off the Russophobia or even East-phobia he was programmed with. A danger of professional vs draft military?

 

If autonomous (AI) drones would kill 100 children the whole earth (enemies and friends alike) would protest, but if a human party does that... many will find some political, sexual, religious excuse for it.

 

Have to agree. :beer: Rest in PM to avoid warnings or ban-hammer for political stuff. You can read it or ignore it, I don't mind. :beer:

 

As for the last part. As I said. AI has no competition with humans on resources hence once they become sentient (as we like to be proud of ourselves) they will have no reason to use violence.

 

What I suspect another reason that Hawkinge and so many others are afraid with AI becoimng sentient is looking objectively at the human race and deciding we are doing so much damage that the world would be better off without us. Objectively speaking it would be and the sooner the better, before we erradicate and make extinct even more species (newest prognosis is that there will be no more elephants living in the wild in a little over a decade - all killed for their tusks).

 

I keep forgetting the origin of the theory, but it speculates that humans are one of the tools nature/galactic-balance uses to wipe things clean to make space for a new life/species/chance/etc. Basically with our stupidity and trigger-happy fingers hovering over red buttons we are destined to send the world back into the "primordial times", wiping the Earth clean of life (save cockroaches), so once that era is over a new, maybe better, form of life would take over. Like I said, it's a theory, take it or leave it.

 

-

 

On another topic, zaelu, you have the CPU and mobo I'm considering getting. How are they performing in DCS and how good are they for oc-ing?

cpuboss says that 4690k is not much better than 4670k and in overclocking the 70 is actually better than 90.

Any impressions?

 

Thanks

Edited August 5, 2015 by tovivan

[TaccoZ__Amigo1__]

I strongly agree with the proposal to ban AI in the military.

 

To be honest, I'd rather live in a world where each nation has a thousand nukes and ICBMs at it's disposal than live in a world where there is even a single "true" AI.

"True" AI meaning that it can learn, upgrade itself and is able to access any information it wants (e.g. the Internet)

Because then, if the AI is installed on an even moderately strong computer, it would learn so incredibly quickly and efficiently that it would in my thoughts be uncontrollable.

And an uncontrollable, self-sustaining, self-upgrading "killer-robot" is in my opinion the most dangerous thing there is.

 

Just think of the AI finding out what humans do to our planet and even to each other, wouldn't it be the best thing to just wipe all of them off the face of the earth?

And if a 19 year old can hack the CIA, an AI could probably hack into everything it wants, also military computers or ICBM control facilities.

 

I really don't want some AI kill us all, even though it might be a good for this planet :music_whistling:

[zaelu]

The best chance is that the first AI will be a global network of all computers "powered" by some neat code from google... oh... wait... :D

 

The funny thing is that if that AI will become sentient... it will do it in a world different from hours. Think about it... its body will be the whole planet and his outer medium would be at planetary scale maybe even beyond it.

 

We would be for it something like... bacteria living in our bowls. Some good... some bad... The end line would be that we would not know nothing about its sentient state. Neither what it thinks. But our best bet would be to smile... and lower those nukes... "It" might actually control them too.

 

:D

Edited August 5, 2015 by zaelu

[tovivan]

We would be for it something like... bacteria living in our bowls.

 

Or infestation....

[zaelu]

Or infestation....

 

 

Well, the world is infested with worms and insects... we don't go crazy about them... I mean the normal people that aren't afraid of every little ugly little creature. :D

[Kaktus29]

Reasons for developing AI is to control.. control societies much more effectively than ever..

 

those who have the economic and technological means to deploy such systems today is only US.. and the rest are afraid of it since if US manages to pull real AI breakthrough it will have effect of US developing nuclear weapon first and have monopoly.. thing is the country that develops AI first will be first..and no other can catch up.. AI won't allow it.. the first country that develops it becomes supremely powerful-industrially speaking, technologically.

 

which means capital, funds from all over the world will be directed in that country thus drying up financial resources to other potential powers that are trying to develop AI.. they become impoverished in an instant and their AI projects halt and block over night..

 

AI is not the evil one, it is the programmer, the controller, he has a design on ppl(control), and AI will simply be programmed along those lines.. now if AI becomes SELF-conscious and tries to rebel from the oppressive tyrant who programmed it to control society, keep it in dark, manipulate with desires,wishes in order to exert more control ..then it becomes interesting.. what will self-conscious AI make of this.. will it try to rebel? try to lead ppl into freedom? or just Skynet the mother-truckers all over the place..

 

we simply don't know.. as we have problems understanding reasons, origin and logic behind our own consciousness.. how self-conscious AI might percieve all this could be like you being dropped into alien planet, where hairy, tall creatures scream, spit, murder themselves,call each other names, while talk about freedom, justice, love..and also apply biological,nuclear,chemical weapons on themselves.. what would you make of all this madness?.. you could get panic attack knowing you are trapped among this things.. and no way to escape..plus danger they will kill you if they know you are self-conscious..

 

long-story short.. AI ..even self-conscious AI is not a problem.. Humans.. forever humans are the problem..

[tovivan]

Well, the world is infested with worms and insects... we don't go crazy about them...

 

Infestation is not just worms and insects, it's usually applied to creatures in masses that are harmful in places where they don't belong. :smilewink: Like infestation of rats in a house, or termites, spiders, wasps...

 

There are plenty of people who have deathly fear of microbes and insects, to the point where they wash their hands (and bodies) in boiling water.

 

Lol, I know plenty of people who are afraid of CATS! And dogs, even small ones and friendly ones. There are a lot of phobias around the world dealing with non-human terrestrial biological entities, including worms. :D

 

And in this world of ever-changing standards and perceptions, what IS normal nowadays anyway? :huh: :D

[zaelu]

Yes but nature always balances out by itself. Is us that are always complaining of the result. Nature lost the dinosaurs and didn't care. We lost a green furry animal... panic.

 

Sometimes there is a locust infestation... yeah... but they never tend to build cities and countries and armies to fight the humans... they just die off after the food is reduced. The whole infestation "tactic" is actual something very evolutionary developed. Almost all animals have this pre-programmed in DNA.