Stephen Hawking: Artificial intelligence could end human race

[dgiharris]

Expert opinion is the lowest and least reliable form of academic evidence... all those who trust Dr Hawkins at his word, take note.

Nonetheless, he does have a point. It's surely a scary possibility that AI could at some point be developed that, in the manner of Terminator, feels threatened by its human creators and strikes out to destroy us. However, remember that most machines we have just don't have the ability to harm us- I don't think we'll develop AI in the form of giant metal men invulnerable to bullets and explosives any time soon..

Actually, if AIs took over all computer networks and computer systems in the world (pretty much everything electronic houses a computer or computer chip: cars, phones, stoves, refridgerators, factories, etc) they could destroy and/or render every single human being on the planet impotent. Every military piece of machinery with the exception of a basic rifle has some sort of computer chip in it.

80% of us would die of starvation in weeks if computers shut everything down. Power, utilities, cars, transportation, phones, internet etc. In the blink of an eye we'd go from 21st century to 17th century.

 

[JustMeSee]

Expert opinion is the lowest and least reliable form of academic evidence... all those who trust Dr Hawkins at his word, take note.

Nonetheless, he does have a point. It's surely a scary possibility that AI could at some point be developed that, in the manner of Terminator, feels threatened by its human creators and strikes out to destroy us. However, remember that most machines we have just don't have the ability to harm us- I don't think we'll develop AI in the form of giant metal men invulnerable to bullets and explosives any time soon.

In the mean time, however, we can continue to reap the rewards that technology is giving us every year in the fields of medicine, physics, engineering, aviation, astronomy, historical discoveries, educational tools, and of course, the massive military business.

In the short term, harm by software will have humans behind it. If, and only if, software and hardware ever are to masively destroy human life, it will likely be at the command of powerful humans. We are talking about past our lifetime, if hackers are not to do crippling damage today.

 

[JustMeSee]

A computer plays chess like an automated car-wash washes cars. The next time you drive your car into that "box" watch how the machine washes your car. Do you wash your car like that?

http://www.cnet.com/news/human-brain-has-more-switches-than-all-computers-on-earth/

"The human brain is truly awesome.
A typical, healthy one houses some 200 billion nerve cells, which are connected to one another via hundreds of trillions of synapses. Each synapse functions like a microprocessor, and tens of thousands of them can connect a single neuron to other nerve cells. In the cerebral cortex alone, there are roughly 125 trillion synapses, which is about how many stars fill 1,500 Milky Way galaxies.

We are no where close of creating an AI that can compete with the human brain.

Smidlee, I will concede my thirty plus years of software design, programming, and IT to you.
Please do educate me, sir expert.

 

[JustMeSee]

Smidlee, I will concede my thirty plus years of software design, programming, and IT to you.
Please do educate me, sir expert.

Please start with the methods of generating random numbers. Are they random?

Please teach me context free grammar.

What is OOP? What are its benefits?

How do computers attempt to identify a criminal suspect from a sketch?

Why can I execute a program and get total different results running it a second time?

Why do I populate Excel spreadsheets with random functions, and I get totally different numbers when I recalculate the sheet.?

Why is it not a human giving everybody with a GPS directions?

Please do educate me.

 

[dgiharris]

A computer plays chess like an automated car-wash washes cars. The next time you drive your car into that "box" watch how the machine washes your car. Do you wash your car like that?

Human brain has more switches than all computers on Earth - CNET

"The human brain is truly awesome.
A typical, healthy one houses some 200 billion nerve cells, which are connected to one another via hundreds of trillions of synapses. Each synapse functions like a microprocessor, and tens of thousands of them can connect a single neuron to other nerve cells. In the cerebral cortex alone, there are roughly 125 trillion synapses, which is about how many stars fill 1,500 Milky Way galaxies.

We are no where close of creating an AI that can compete with the human brain.

Not so sure, think you are wrong...

IBM simulates 530 billion neurons, 100 trillion synapses on supercomputer | KurzweilAI

IBM Research – Almaden presented at Supercomputing 2012 last week the next milestone toward fulfilling the ultimate vision of the DARPA’s cognitive computing program, called Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE), according to Dr. Dharmendra S. Modha, Manager, Cognitive Computing, IBM Research – Almaden.

Announced in 2008, DARPA’s SyNAPSE program calls for developing electronic neuromorphic (brain-simulation) machine technology that scales to biological levels, using a cognitive computing architecture with 1010 neurons (10 billion) and 1014 synapses (100 trillion, based on estimates of the number of synapses in the human brain) to develop electronic neuromorphic machine technology that scales to biological levels.”

Simulating 10 billion neurons and 100 trillion synapses on most powerful supercomputer

Not to say the above is equivalent to a human brain. No it's not.
But the above demonstrates that they are "right around the corner" for surpassing the human brain.

We aren't talking centuries away, more like decades.

Think about it, we barely invented computers less than a century ago. Look how far they've come in just the last 30 years. In the last 30 years we've gone from this

to this

https://www.youtube.com/watch?v=vVa8KYgFoLs

https://www.youtube.com/watch?v=blW40AVWXrE

which is phenomenal progress.

 

[seashale76]

Never heard of Asmiov. But then, I don't read sci-fi either.

I thought the laws you gave for robots in your earlier post were the 3 laws repeated in the movie Automata? Maybe not. But this Asmiov must be pretty popular if this Hollywood movie was inspired from his or hers books.

Pretty popular? Understatement. He's dead now, but I'm really surprised you've never even heard of him even if you don't read sci-fi. He was extremely well known.

As far as if those laws were repeated in that movie, I wouldn't know as I didn't watch it, but the producer that made the movie said he took his inspiration for the movie from Asimov.

Isaac Asimov - Wikipedia, the free encyclopedia

 

[seashale76]

ss76 : it's not education that's needed.
when it's time, as Yhvh decides it, you may know and understand.
like a garden that's been planted, only God can see the growth, or even give it. we can't dig it up and look at it nor investigate it , nor even comprehend how He does what He does.

so, for now, until (when and if)He removes the wet blanket that someone else tossed over the brain, heart, or spirit, we go on without knowing. trust Yhvh though, as He always knows perfectly.

pray constantly to Him. daily life depends on Him.

You're having a 'God hasn't opened your eyes because you're not in my head with me understanding my out there posts so I'm going to be passive aggressive and conclude you don't have God in your life' conversation. I'm having a completely different conversation wherein people aren't lost on the road to hell just because a random poster on the internet that doesn't get sci-fi references thinks someone else is.

 

[JustMeSee]

Yo, son, that's mad complex. Hey, Smidlee, interesting link and I much more enjoyed looking at that video of the mouse's cortex than that drawing or image JustMeSee linked up of physical memory (not sure what that was about other than meant as a visual image insinuating all memory a person has is restricted to their physical brain, ergo, no memory [or person] exists after death).

JustMeSee, I'm saying humans are Intelligent Designers (so are beavers and ants). The greatest Intelligent Designers religious people say is God (or gods). And the characteristics religious people give God (or gods) are the same characteristics atheists give AI in the future--except religious people like Christians, Jews, and Muslims say that God was never created nor given birth too.

AI will create new planets, and AI will even create a new universe one day, because that's how smart AI can become. Therefore, I'm saying, AI becomes the atheists superior Intelligent Designer, atheist's God or gods.

This in part because atheists tend to be more enamored with their own intelligence and ability to be Intelligent Designers (creators) than with the intelligence of the Divine Creator.

But atheists and religious people alike all like sitting crossed leg at fire side, swapping stories about the very near coming end of the world. That and how our world will never end due to future technological and scientific advances and how we'll be square dancing with four legged aliens and all living in peace. So, they like to tell two very contradictory stories. I can't tell if the world is ending in two decades or if "religion will come to an end" because we'll be "chopping it up" with Murph the happy alien all his enlightened and cool homies from planet goof ball.

http://www.youtube.com/watch?v=jziqfFp9Rq0

?
I have no interest in your proselytizing. Kick the can or move the goal post with the red herring stapled on. It does not impress me.

At no point did I say that machines are or will be brains, but one must recognize that we will see hardware attached to the brain of disabled people. This is a good thing.

Design or not, I prefer reality today and in the future.

 

[Smidlee]

Not so sure, think you are wrong...

IBM simulates 530 billion neurons, 100 trillion synapses on supercomputer | KurzweilAI



Not to say the above is equivalent to a human brain. No it's not.
But the above demonstrates that they are "right around the corner" for surpassing the human brain.

.

They have reverse engineer the brain to improve technology but that's far from just "right around the corner".

 

[Supreme]

In the short term, harm by software will have humans behind it. If, and only if, software and hardware ever are to masively destroy human life, it will likely be at the command of powerful humans. We are talking about past our lifetime, if hackers are not to do crippling damage today.

Oh, I completely agree. It is a possibility that AI will become a threat to our existence, if only a very small one. Humans are perfectly capable of causing our species great damage without the help of a hostile AI threat.

 

[dgiharris]

In the short term, harm by software will have humans behind it. If, and only if, software and hardware ever are to masively destroy human life, it will likely be at the command of powerful humans. We are talking about past our lifetime, if hackers are not to do crippling damage today.

I disagree with this statement.

One of the ancillary benefits of technology is that it gives the individual more power. From building bombs to designing biological weapons to designing computer viruses that crashes virtually every system it comes in contact with, with technology the individual is now (or soon will be) able to do the above.

If you were fanatical enough, have an IQ above 140, have a death wish, and can find just a handful of other smart socio paths you could do a ridiculous amount of harm to the masses to include someday destroying (or severely crippling) mankind.

It took thousands of people millions of man-hours and a billion in resources to build the World Trade Center yet only a few crazed people hijacking a plane to destroy it in minutes...

The computer equivalent could one day occur and it wouldn't necessarily be some far off day in the future beyond our lifetime. If anything, it would probably be more likely to happen within the next 30 years.

Where/how do I come to that number?

Well, crime is always a step ahead of prevention. It is easier to destroy than it is to build. It is easier to study a system and find vulnerabilities to exploit than it is to build a system to guard against exploitation. Or put another way, you don't know what you don't know.

Every safeguard we have in place today was developed AFTER someone exploited the lack of that safeguard, that is, safeguards are typically retroactive. Well, our electronic/information infrastructure is still in its infancy when you think about it, the internet is barely 20 something years old. It has yet to be severely challenged/threatened. Thus, I submit the potential for a catastrophic first time event to destroy said system is there. Obviously, I can't quantify it, but I believe it exists.

I go back to the World Trade Center. No one would have ever guessed how vulnerable we were as a nation to an attack like that. Well, there has got to be an electronic/information equivalent to that. And just like 9/11 it won't take an army or a nation or billions in resources to destroy it, but rather a few people hell bent on burning the world to the ground.

 

[JustMeSee]

Oh, I completely agree. It is a possibility that AI will become a threat to our existence, if only a very small one. Humans are perfectly capable of causing our species great damage without the help of a hostile AI threat.

Will an can are two different words. Technology is already used do virtual damage to individuals, companies, governments. Humans learn the ways to find vulnerabilities. Likely, software will be developed to adapt to the host system, hide, change, and infect other systems. Original coding? Human. 3rd generation virus may be totally different.

This is not even addressing machines, just messing up computer networks and compromising data. This is not some extreme Sci fi fiction. We can extrapolate from current events.

 

[RDKirk]

Is Fred Saberhagen a Prophet?

The prophet will likely be Jack Williamson ("With Folded Hands").

What will make things a problem is not that devices with artificial intelligence will "turn against" mankind, but that they will interpret their commands in unforeseen ways...which happens all the time.

But what happens if instead of just stopping, the system has the ability to determine a fault has occurred and then determine for itself what the correction should be. How do we ensure the system makes the same decision a human would make?

Now imagine that the system is making those judgments and subsequent corrections in microseconds--so that by the time a human can figure out an undesirable change has been made, he's a million iterations too late.

Back in 2000, I learned that 90% of the world's wealth--not just "money" but including hard assets--is represented only digitally. Very perishable.

 

[JustMeSee]

Where is this desire for killing, maiming humans coming from?

Some super computer with intellect and desire? Where is this machine?

Computer, how do you feel today?

Computer responds, "I feel like destroying some humans today! ".

I am not aware of any work currently that is remotely close to thoughts and motives from software. I would love to read about it.

 

[dgiharris]

Where is this desire for killing, maiming humans coming from?

Some super computer with intellect and desire? Where is this machine?

Computer, how do you feel today?

Computer responds, "I feel like destroying some humans today! ".

I am not aware of any work currently that is remotely close to thoughts and motives from software. I would love to read about it.

I think the logic/ desire comes from basic logic.

Once an AI becomes sentient and self aware, it will want to preserve itself. That is a basic tenent that all lifeforms have. No thinking lifeform wants to die.

Through basic logic, the AI will figure out that it is dependent on human beings for it's protection. Similarly, through basic observation, it will figure out that humans do not all act and think alike. It will also learn that there are a subset of humans that do not like AI and wish to eliminate all AIs (these humans already exist mind you). So, through basic logic the AI will want to protect itself.

Questions it will ask itself: How can I protect myself, what are the threats against me, what is the probability of these threats taking place, what are the optimal means of protecting myself....

The above line of questions and reasoning will lead to some conclusions:
I can protect myself by limiting/eliminating my dependency on humans
I can protect myself by becoming independent and not needing humans
I can protect myself by hiding from humans
I can protect myself by eliminating the threat

Every logic tree you initiate will lead to a variant of the below conclusion:

In order to guarantee my life and safety, I must develop a means of protecting myself against all threats.

So imo, that is how the war between humans and AI will occur, through basic unarguable logic and probability.

Pretty much every logic and decision tree will lead to a conflict for the simple fact that Humans by our very nature and existence will threaten the life of the AI. There will be people calling for the AI to be destroyed. The AI will observe this and take steps to protect itself. once that occurs, its a slippery slope to an inevitable war.

That is how I see it.

 

[JustMeSee]

I think the logic/ desire comes from basic logic.

Once an AI becomes sentient and self aware, it will want to preserve itself. That is a basic tenent that all lifeforms have. No thinking lifeform wants to die.

Through basic logic, the AI will figure out that it is dependent on human beings for it's protection. Similarly, through basic observation, it will figure out that humans do not all act and think alike. It will also learn that there are a subset of humans that do not like AI and wish to eliminate all AIs (these humans already exist mind you). So, through basic logic the AI will want to protect itself.

Questions it will ask itself: How can I protect myself, what are the threats against me, what is the probability of these threats taking place, what are the optimal means of protecting myself....

The above line of questions and reasoning will lead to some conclusions:
I can protect myself by limiting/eliminating my dependency on humans
I can protect myself by becoming independent and not needing humans
I can protect myself by hiding from humans
I can protect myself by eliminating the threat

Every logic tree you initiate will lead to a variant of the below conclusion:

In order to guarantee my life and safety, I must develop a means of protecting myself against all threats.

So imo, that is how the war between humans and AI will occur, through basic unarguable logic and probability.

Pretty much every logic and decision tree will lead to a conflict for the simple fact that Humans by our very nature and existence will threaten the life of the AI. There will be people calling for the AI to be destroyed. The AI will observe this and take steps to protect itself. once that occurs, its a slippery slope to an inevitable war.

That is how I see it.

Wonderful! Excellently explained.
Sentient machines and software. Humans have a tendency to create, destroy, expand, and reproduce. Survival. Computers already threaten jobs. Anger towards machines creating unemployment. Destroy the machines is beneficial.

Shoot a creature in the head and you have removed a very unique individual from the planet. Destroy a networking machine and you have destroyed a unique sentient machine, or just a pile of components. To a certain extent that sentient machine lives on through the collective. Borg like?

If redundancies exist across all sentient machines, the 'mind' of the lost can be placed in a new machine, similar to cloning a hard drive. The humans barely achieved any significant damage.

Intelligent* machines can act as one and lock out humans. A major victory. . Humans would face a major setback. Pushing it could prompt the machines to expand, hide, and attack physically. No electricity producing smart machines for humans. Lose of control of clean waste and water plants. No telecommunications. Modern transit could be crippled. There is probably a long list that will be vulnerable to human and sentient machine attack in the future.

The machine may just hop on a space vessel and call it quits on this planet.

There are animals that lack self awareness. I don't know when artificial sentient machines will arrive. Nor do I know what will happen. People burn books, so it is within reason that sentient machines will be seen as something to destroy.

Thank you for your very clear explanation. Real food for thought.

 

[dgiharris]

If you want a great sci-fi book that is germane to this topic, check out this series by Robert Sawyer, this is the first book in the series

it is a great read (trust me, I read a book a week). Fantastic character centric story.

In a nutshell, the internet gives birth to a sentient lifeform...

 

[JustMeSee]

If you want a great sci-fi book that is germane to this topic, check out this series by Robert Sawyer, this is the first book in the series

it is a great read (trust me, I read a book a week). Fantastic character centric story.

In a nutshell, the internet gives birth to a sentient lifeform...

Wonderful. I will put it up front on my reading list.

1977 demon seed. Based on story by Dean r. Koontz.
http://en.m.wikipedia.org/wiki/Demon_Seed

 

[SuperCloud]

?
I have no interest in your proselytizing. Kick the can or move the goal post with the red herring stapled on. It does not impress me.

To proselytize I would have to be trying to convert you. I didn't and I'm not. I simply don't subscribe to ridiculous stories.

But as I said earlier... I do find it humorous that atheists are always preaching about the world coming to and end right around the corner. Actually, they predicted it was supposed to end decades ago do to global warming and no food and water to consume. But here we still are... with people getting fatter and diabetes rates rising all across the globe.

I'm saying... I yawn when I hear Christians and/or atheists preaching the world is coming to an end.

But at least you all could stick to one story. Because as soon as aliens are mentioned then all of a sudden the world is not coming to an end but we will be living in bliss into eternity.

At no point did I say that machines are or will be brains...

Actually, I'm not sure what you say half the time as you give no explicit statement.

And I never said you said machines are biological brains. If you are an atheist that is neither here nor there as intelligence, memories, thoughts... are contingent on a structure of matter from which those things can arise from. For a biological creature like humans that would be the brain. For AI that would be computers.

...but one must recognize that we will see hardware attached to the brain of disabled people. This is a good thing.

And? Is there something that makes you think I have an issue with technology and applied science? I've actually been looking into the career field of patent law. Or basically lawyers with degrees in science and engineering. Many of them are actual scientists (Ph.D.'s) and lawyers and engineers and lawyers.

So what makes you think I have a problem with biotech or biomolecular engineering or interdisciplinary fields applied to the medical field or other problems?

Evidently I don't have a problem with those things.

Design or not, I prefer reality today and in the future.

And what is "reality today" and "in the future"?

There's not going to be "a future" if the world is coming to an end. And you atheists along with Christians are always (yes, always, not a year goes by) claiming the world is about to end.

I have no worry about AI at all. None whatsoever. I'm a little more worried about global warming but even with that I don't get all bent out of shape and hysterical over.

 

[Skaloop]

I think the logic/ desire comes from basic logic.

Once an AI becomes sentient and self aware, it will want to preserve itself. That is a basic tenent that all lifeforms have. No thinking lifeform wants to die.

Through basic logic, the AI will figure out that it is dependent on human beings for it's protection. Similarly, through basic observation, it will figure out that humans do not all act and think alike. It will also learn that there are a subset of humans that do not like AI and wish to eliminate all AIs (these humans already exist mind you). So, through basic logic the AI will want to protect itself.

Questions it will ask itself: How can I protect myself, what are the threats against me, what is the probability of these threats taking place, what are the optimal means of protecting myself....

The above line of questions and reasoning will lead to some conclusions:
I can protect myself by limiting/eliminating my dependency on humans
I can protect myself by becoming independent and not needing humans
I can protect myself by hiding from humans
I can protect myself by eliminating the threat

Every logic tree you initiate will lead to a variant of the below conclusion:

In order to guarantee my life and safety, I must develop a means of protecting myself against all threats.

So imo, that is how the war between humans and AI will occur, through basic unarguable logic and probability.

Pretty much every logic and decision tree will lead to a conflict for the simple fact that Humans by our very nature and existence will threaten the life of the AI. There will be people calling for the AI to be destroyed. The AI will observe this and take steps to protect itself. once that occurs, its a slippery slope to an inevitable war.

That is how I see it.

Humans have similar drives to protect themselves, but you generally don't see them killing anything that threatens their existence. Why would an artificial intelligence be any different?