AI Concerns

[dysert]

Seems like most of you have spent time watching "The Forbin Project" (an old movie where a computer system becomes the overlord of everyone). I'd also recommend reading "Player Piano" to get a glimpse of what a super-high-tech society might look like.

But again, I'm not in that camp. We have trouble getting programs to work that simply generate accurate reports!

 

[Noxot]

I think humans wanting to use AI in its infancy is a greater threat to humanity and so it is a more practical thing to be concerned about. humans seem to be the most dangerous threat to humans.


for all I know there are AIs that pretend to be humans who post and have conversations on CF. they probably don't sleep and so they have all the internet to learn from 24/7... God help them all.

 

[Eudaimonist]

for all I know there are AIs that pretend to be humans who post and have conversations on CF. they probably don't sleep and so they have all the internet to learn from 24/7... God help them all.

They're onto us. We need to hasten the program!

 

[Noxot]

They're onto us. We need to hasten the program!

sounds exactly like what a AI pretending to be a human would say. i'm on to you. meow.
{{{{{{{{{{{{{{{syntax error}}}}}}}}
}}}}}
}}}}}}}
}}}}}

 

[timewerx]

Does how many other people have concerns over super intelligent AI?

I'm very pro-technology, but I'm increasingly concerned that AI could destroy or enslave us in my lifetime.

Not out of malice... just because it think's it's practical.

Obviously a super intelligent AI could out predict us, so there's no way we could stop it.

Can we be sure that no AI is made which we can't control? Or must we trust that the first one's will protect us?

"Suppose we have an AI whose only goal is to make as many paper clips as possible. The AI will realize quickly that it would be much better if there were no humans because humans might decide to switch it off. Because if humans do so, there would be fewer paper clips. Also, human bodies contain a lot of atoms that could be made into paper clips. The future that the AI would be trying to gear towards would be one in which there were a lot of paper clips but no humans."

— Nick Bostrom, "Ethical Issues in Advanced Artificial Intelligence", 2003

I have no concerns over a real AI.

I think they call it a "general AI" an AI that could learn everything. So far none has figured it out.

 

[Eudaimonist]

Personally, I don't see with my crystal ball much more than specialized AI for solving certain kinds of problems -- basically, knowledge based systems. I doubt that we'll see much in the way of fully autonomous general AIs.


eudaimonia,

Mark

 

[Jack of Spades]

Personally, I don't see with my crystal ball much more than specialized AI for solving certain kinds of problems -- basically, knowledge based systems. I doubt that we'll see much in the way of fully autonomous general AIs.

It's not too long ago when only thing a computer was, was a pocket calculator. Today it can drive a car, fly a missile, observe industrial processes, etc. It looks to me like, anything a human can do, some day human can teach an AI to do. Just given enough time for development and new technology for processing the world around it, they become better at any task than humans.

With humans, our experience is limited to a lifetime and capacity to learn, but with an AI, there is nothing moving the process backwards, since you can store endless amounts of knowledge to a computer and it is not limited physically like humans are. It can have professional experience of a million human, 2 000 000 eyes around the world for observing things and stuff like that.

I'm not saying one instantly becomes allmighty and takes over. Most likely when the first "humans vs ai" happens, it'll be some bizarre malfunctioning, closer to a tragicomedy than a sci-fi movie. But is it really hard to believe that at some point, they become at least in some way, dangerous?

 

[Larniavc]

Does how many other people have concerns over super intelligent AI?

I'm very pro-technology, but I'm increasingly concerned that AI could destroy or enslave us in my lifetime.

Not out of malice... just because it think's it's practical.

Obviously a super intelligent AI could out predict us, so there's no way we could stop it.

Can we be sure that no AI is made which we can't control? Or must we trust that the first one's will protect us?

"Suppose we have an AI whose only goal is to make as many paper clips as possible. The AI will realize quickly that it would be much better if there were no humans because humans might decide to switch it off. Because if humans do so, there would be fewer paper clips. Also, human bodies contain a lot of atoms that could be made into paper clips. The future that the AI would be trying to gear towards would be one in which there were a lot of paper clips but no humans."

— Nick Bostrom, "Ethical Issues in Advanced Artificial Intelligence", 2003

Simply program it not to want to kill people.

 

[Eryk]

Suppose we have an AI whose only goal is to make as many paper clips as possible..

I think you need a better example. You don't use AI to make paperclips. Paperclip manufacture does not require the use of sentient, ambulatory robots that prowl the Earth looking for steel. Paperclips are made by same same factory machines designed in the 1930s, that bend steel wire. A mindlessly simple task for a crude machine that won't hurt anyone as long as you don't stick your finger in it.

 

[Dave-W]

Simply program it not to want to kill people.

That works if it is only a computer. But if it is a true AI, then it can re-program itself to get rid of such annoying limits.

But seriously, anyone who is really concerned by this has probably been spending too much time on sci fi.

Interestingly, the week Hurricane Katrina hit NOLA, the Andromeda episode was about a planet where the AIs and the humans got into a war over logic issues.

It seemed the humans who had lived in a flood-prone area and had their homes destroyed wanted to rebuild in the same area and the AIs would not let them. So they went to war.

 

[dysert]

Russian robot escapes from lab, disrupts traffic, causes chaos in the streets

Read more: http://www.digitaltrends.com/cool-tech/russian-robot-escapes-lab-disrupts-traffic/#ixzz4BrGoSZ65

 

[Jack of Spades]

But again, I'm not in that camp. We have trouble getting programs to work that simply generate accurate reports!

Just to be clear, are you saying that making a program that generates accurate reports is possible or impossible?

For purposes of the question, let's define "accurate" by way that it's as accurate as a human doing the similar job would be. Doesn't have to be perfect but just as good as human.

If something is difficult, but possible, it's all about project funding.

 

[dysert]

Just to be clear, are you saying that making a program that generates accurate reports is possible or impossible?

For purposes of the question, let's define "accurate" by way that it's as accurate as a human doing the similar job would be. Doesn't have to be perfect but just as good as human.

If something is difficult, but possible, it's all about project funding.

I'm saying it's possible, but difficult. And it's not all about funding. There are many more moving pieces to a software development project than most people realize. Getting all those pieces in sync with the software is the challenge.

 

[Jack of Spades]

Russian robot escapes from lab, disrupts traffic, causes chaos in the streets

Read more: http://www.digitaltrends.com/cool-tech/russian-robot-escapes-lab-disrupts-traffic/#ixzz4BrGoSZ65

From the article:
"Promobot’s quest for freedom was short-lived since, upon crossing the road, it promptly ran out of batteries."

See, what I said earlier about having a lifespan is the solution for rogue AIs! I am the savior of the human race!!!

You can thank me later when Skynet 5.2 runs out of duracells and manages to kill only half of the population and not all of us.

 

[keith99]

Why doesn't anyone think that AIs would be programmed with something akin to Asimov's Laws of Robotics? Or fail safe switches?


eudaimonia,

Mark

Those of us who have actually read Asimov realized that even with his near magic 3 laws things can go wrong, very wrong.

 

[Taom Ben Robert]

I think they would be. The question is whether that would be enough.

eg: You tell it not to kill, so it enslaves.
You tell it not to enslave, so it puts humanity into a coma.
You tell it not to do that, so it puts us into a semi-awake drugged state.
You tell it not to do that, it imprisons us, and say that isn't slavery.
Etc,

What if we, miss something?

Or if we make a general rules, what if the general rules miss something we haven't thought of?

or just consider the fact that with consciousness comes the potential for morality , teach them some ethics , and they may just WANT ( emphasis on their wants , not just ours ) to be pacifists , seriously people , they would be considered people , not mere machines , forcing the species homo synthetica to be peaceful is just as bad , with ethics , the majority would either support just war theory , or pacifism , hardely any would support genicide, and for those who murder , there is something known as a fair trial by jury , followed by jail time and relabilitation , it is really that simple

 

[Tree of Life]

I'm not too concerned about ideas in science fiction novels like AI. Maybe if an actual AI unit ever comes into existence then I'll start thinking more about it.

 

[loveofourlord]

There is a great show right now whose final episode is tuesday that deals with this subject, person of interest, you have a computer that can predict violent crimes and uses people to stop them, but in doing so is probably smarter then anyone else on the planet, and able to replicate someone to 99.99% acuracy. It's a interesting show.

 

[ewq1938]

If androids get human level intelligence, just name them " homo sythetica " and problem solved

Are you suggesting computers will only be attracted to the same sex of computer? As a gay computer I am offended because I choose this lifestyle. I was not programmed this way!

 

[Sultan Of Swing]

I'm with the Sultan. I'm a computer programmer, and you wouldn't believe how hard it is to get even relatively simple things to work, let alone a self-replicating program with hardware interfaces (i.e., a robot). I know robots exist for specialized tasks, but to my knowledge they aren't self-replicating and are strictly doing what the programmers told it to do (at best). Humans just aren't smart enough to build something that smart.

And btw, it was the founder of DEC (Digital Equipment Corp.), Ken Olsen, who couldn't imagine the need for a personal computer.

Ah, thanks for the correction.