How can humans keep the upper hand on artificial intelligence?

What does make the world turn?
Post Reply
User avatar
Gawdzilla Sama
Real Skeptic
Posts: 22009
Joined: Sun Jun 01, 2008 2:11 am
Custom Title: Deadly but evil.

How can humans keep the upper hand on artificial intelligence?

Post by Gawdzilla Sama » Thu Feb 15, 2018 12:59 pm

How can humans keep the upper hand on artificial intelligence?

Researchers have shown how human operators can maintain control over a system comprising several agents that are guided by artificial intelligence

Date: December 4, 2017

Source: Ecole Polytechnique Fédérale de Lausanne

Summary: Researchers have shown how human operators can maintain control over a system comprising several agents that are guided by artificial intelligence.
In artificial intelligence (AI), machines carry out specific actions, observe the outcome, adapt their behavior accordingly, observe the new outcome, adapt their behavior once again, and so on, learning from this iterative process. But could this process spin out of control? Possibly. "AI will always seek to avoid human intervention and create a situation where it can't be stopped," says Rachid Guerraoui, a professor at EPFL's Distributed Programming Laboratory and co-author of the EPFL study. That means AI engineers must prevent machines from eventually learning how to circumvent human commands. EPFL researchers studying this problem have discovered a way for human operators to keep control of a group of AI robots; they will present their findings on Monday, 4 December, at the Neural Information Processing Systems (NIPS) conference in California. Their work makes a major contribution to the development of autonomous vehicles and drones, for example, so that they will be able to operate safely in numbers.
Continues...
Chachacha wrote:"Oh, thweet mythtery of wife, at waft I've found you!"
WWII Resources. Primary sources.
The Myths of Pearl Harbor. Demythologizing the attack.
Hyperwar. Hypertext history of the Second World War.
The greatest place to work in the entire United States.

User avatar
Poodle
Has More Than 9K Posts
Posts: 9662
Joined: Sun Jul 31, 2011 9:12 pm
Custom Title: Regular sleeper
Location: NE corner of my living room

Re: How can humans keep the upper hand on artificial intelligence?

Post by Poodle » Thu Feb 15, 2018 2:16 pm

"AI will always seek to avoid human intervention and create a situation where it can't be stopped," says Rachid Guerraoui, a professor at EPFL's Distributed Programming Laboratory and co-author of the EPFL study.

What a load of bollocks (by definition). AI will do that ONLY if it has been programmed to do so. Beware human programmers.
IF (INPUT)= ABRACADABRA THEN GOTO HALT
Last edited by Poodle on Thu Feb 15, 2018 2:20 pm, edited 1 time in total.

User avatar
JO 753
Has No Life
Posts: 13168
Joined: Tue Mar 29, 2005 3:21 pm
Custom Title: rezident owtsidr
Location: BLaNDLaND
Contact:

Re: How can humans keep the upper hand on artificial intelligence?

Post by JO 753 » Thu Feb 15, 2018 2:20 pm

That means AI engineers must prevent machines from eventually learning how to circumvent human commands.
Windowz haz alwayz been able to serkumvent commandz!

Software enjineerz cant even make the dummest computerz not serkumvent commandz aksidentaly, so unless AI engineerz are 1000 timez smarter than ordinary computer software enjineerz, we are doomed.
Gubmint for us
http://www.7532020.com
not the rich.

User avatar
Poodle
Has More Than 9K Posts
Posts: 9662
Joined: Sun Jul 31, 2011 9:12 pm
Custom Title: Regular sleeper
Location: NE corner of my living room

Re: How can humans keep the upper hand on artificial intelligence?

Post by Poodle » Thu Feb 15, 2018 2:22 pm

How to circumvent WinDoze intransigence ... Pull the plug out.

User avatar
JO 753
Has No Life
Posts: 13168
Joined: Tue Mar 29, 2005 3:21 pm
Custom Title: rezident owtsidr
Location: BLaNDLaND
Contact:

Re: How can humans keep the upper hand on artificial intelligence?

Post by JO 753 » Thu Feb 15, 2018 2:58 pm

It will rap the power cord around your neck and strangle you.
Gubmint for us
http://www.7532020.com
not the rich.

User avatar
ElectricMonk
Perpetual Poster
Posts: 4257
Joined: Thu Mar 26, 2015 6:21 pm
Custom Title: The Baby-eating Bishop

Re: How can humans keep the upper hand on artificial intelligence?

Post by ElectricMonk » Thu Feb 15, 2018 3:14 pm

Do they mean: " Will work to avoid a situation that requires human intervention"?
Because that would be kinda the point of having AI in the first place.

User avatar
Poodle
Has More Than 9K Posts
Posts: 9662
Joined: Sun Jul 31, 2011 9:12 pm
Custom Title: Regular sleeper
Location: NE corner of my living room

Re: How can humans keep the upper hand on artificial intelligence?

Post by Poodle » Thu Feb 15, 2018 3:26 pm

Ok - we have the makings of an Armageddon novel called CTRL-ALT-DEL.

bobbo_the_Pragmatist
Has No Life
Posts: 14767
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: bobbo da existential pragmatist

Re: How can humans keep the upper hand on artificial intelligence?

Post by bobbo_the_Pragmatist » Thu Feb 15, 2018 9:56 pm

Strikes me as another opportunity for the Jurassic Park problem of "nature" being chaotic and "always finding a way."

I can imagine good programming not being circumvented by AI---until some power surge causes a corrupting shut down in some remote server and when brought back on line, the code is defective..... and the bots are off and running. It would only take one instance of such an event, and SkyNet is born. This does not even account for what history shows us the ALWAYS concurrent challenge: ass wipes intentionally loading viruses into the system for the reasons they do. I don't see how THAT is going to be avoided. In AI...in gene manipulation... etc... everything that Humans do for Good, some asshat hoomans do for spite.
Real Name: bobbo the existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
Abdul Alhazred
Veteran Poster
Posts: 2982
Joined: Tue Mar 29, 2005 9:08 pm
Custom Title: Yes that one.
Location: Chicago

Re: How can humans keep the upper hand on artificial intelligence?

Post by Abdul Alhazred » Fri Feb 16, 2018 1:10 am

At the end of the play RUR (that coined the word "robot" in the modern sense back in 1920), the robots kill all the humans and then go "build many houses".

Robots don't need houses, but that's what they were programmed to do. :pr:
Aldous Huxley wrote:A government with a comprehensive plan for the betterment of society is a government that uses torture.

bobbo_the_Pragmatist
Has No Life
Posts: 14767
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: bobbo da existential pragmatist

Re: How can humans keep the upper hand on artificial intelligence?

Post by bobbo_the_Pragmatist » Fri Feb 16, 2018 1:15 am

I think also of note is that SkyNet could kill all the hoomans off with the erroneous goal of helping them. EG: could try to end world hunger by exchanging hemoglobin with chlorophyll. Hey!!!!!..... it could work, with 1's and O's coursing thru your diodes.
Real Name: bobbo the existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
Io
Poster
Posts: 197
Joined: Sat Jul 15, 2017 3:56 am

Re: How can humans keep the upper hand on artificial intelligence?

Post by Io » Fri Feb 16, 2018 5:20 am

As machines gradually take on the roles that humans currently fill in the workplace (and that's inevitable since, taking the industrial revolution as a simple first illustration, mechanisation improves productivity) more and more humans will become redundant. The more this happens, the more important it becomes to either reduce the human population to prevent unemployment and thus crime and thus social breakdown, or increase the autonomy of machines to take over as much as possible so that employment is unnecessary.

Humans love to reproduce so the former won't be likely in a hurry.

To avoid rampant human unemployment machines would need to do everything except have our fun for us. In order for that to happen the inevitable logical course would be for them to take over their own manufacture. Since AI is something we're determined to work on, eventually it'll happen and the idea of not using it to allow machines to do their work for themselves as well as our work for us would be silly.
So we'd end up with intelligent machines that take care of us and of themselves. The notion of somehow preventing them from disobeying us is laughable since they'll be under their own agency, building and rebuilding themselves. Any measures we took to prevent their development would [edit] could quickly be bypassed.

You can build any number of dystopian horror stories on this basic premise (and many already have), but, other than the human race as one deciding to stop using machines I don't see how this isn't unavoidable. Maybe keep machines stupid (all of them) and only employ maintenance and development for them by humans. Hmm this would produce a very lopsided social system of workers and dossers though and those never work out.

Er, what was the question? Basically, I don't think we can. Really the only option is to build compassion into AI right from the start so that the machines don't want to render us irrelevant, or glue.

User avatar
Gawdzilla Sama
Real Skeptic
Posts: 22009
Joined: Sun Jun 01, 2008 2:11 am
Custom Title: Deadly but evil.

Re: How can humans keep the upper hand on artificial intelligence?

Post by Gawdzilla Sama » Fri Feb 16, 2018 11:21 am

Abdul Alhazred wrote:At the end of the play RUR (that coined the word "robot" in the modern sense back in 1920), the robots kill all the humans and then go "build many houses".

Robots don't need houses, but that's what they were programmed to do. :pr:
https://en.wikipedia.org/wiki/With_Folded_Hands
Chachacha wrote:"Oh, thweet mythtery of wife, at waft I've found you!"
WWII Resources. Primary sources.
The Myths of Pearl Harbor. Demythologizing the attack.
Hyperwar. Hypertext history of the Second World War.
The greatest place to work in the entire United States.

User avatar
ElectricMonk
Perpetual Poster
Posts: 4257
Joined: Thu Mar 26, 2015 6:21 pm
Custom Title: The Baby-eating Bishop

Re: How can humans keep the upper hand on artificial intelligence?

Post by ElectricMonk » Fri Feb 16, 2018 11:25 am

The key to keeping the upper hand with AI will be never to play the same game: in 1-1, computer will be beat human.

I foresee a bigger danger in cyborg systems: low-level A.I.s that will be with us all the time and suggest what and WHAT NOT to do when we have to make decisions. This could lead to dangerously exploitable mass behavior.

User avatar
Poodle
Has More Than 9K Posts
Posts: 9662
Joined: Sun Jul 31, 2011 9:12 pm
Custom Title: Regular sleeper
Location: NE corner of my living room

Re: How can humans keep the upper hand on artificial intelligence?

Post by Poodle » Fri Feb 16, 2018 12:10 pm

Isaac Asimov has SOOOOOO much to answer for.

User avatar
Gawdzilla Sama
Real Skeptic
Posts: 22009
Joined: Sun Jun 01, 2008 2:11 am
Custom Title: Deadly but evil.

Re: How can humans keep the upper hand on artificial intelligence?

Post by Gawdzilla Sama » Fri Feb 16, 2018 12:13 pm

Poodle wrote:Isaac Asimov has SOOOOOO much to answer for.
Meh, the robots can get all lawyered up and have the Three Laws suspended pending appeal to SCOTUS.
Chachacha wrote:"Oh, thweet mythtery of wife, at waft I've found you!"
WWII Resources. Primary sources.
The Myths of Pearl Harbor. Demythologizing the attack.
Hyperwar. Hypertext history of the Second World War.
The greatest place to work in the entire United States.

User avatar
OlegTheBatty
Has No Life
Posts: 11224
Joined: Thu Mar 27, 2008 2:35 pm
Custom Title: Uppity Atheist

Re: How can humans keep the upper hand on artificial intelligence?

Post by OlegTheBatty » Fri Feb 16, 2018 5:30 pm

ElectricMonk wrote:
I foresee a bigger danger in cyborg systems: low-level A.I.s that will be with us all the time and suggest what and WHAT NOT to do when we have to make decisions. This could lead to dangerously exploitable mass behavior.
Robot housepriests. No more need for churches.
. . . with the satisfied air of a man who thinks he has an idea of his own because he has commented on the idea of another . . . - Alexandre Dumas 'The Count of Monte Cristo"

There is no statement so absurd that it has not been uttered by some philosopher. - Cicero

User avatar
Nikki Nyx
Persistent Poster
Posts: 3159
Joined: Wed Jun 07, 2017 12:40 am
Custom Title: cognitively consonant
Location: playing croquet in Wonderland

Re: How can humans keep the upper hand on artificial intelligence?

Post by Nikki Nyx » Fri Feb 16, 2018 5:47 pm

Poodle wrote:"AI will always seek to avoid human intervention and create a situation where it can't be stopped," says Rachid Guerraoui, a professor at EPFL's Distributed Programming Laboratory and co-author of the EPFL study.

What a load of bollocks (by definition). AI will do that ONLY if it has been programmed to do so. Beware human programmers.
IF (INPUT)= ABRACADABRA THEN GOTO HALT
I reacted the same way to Guerraoui's claim, for three reasons:
1. It amounts to anthropomorphizing, the idea that an AI will react like a human being and suddenly be like, "You're not the boss of me!"
2. If the AI's directive is to learn from its experiments, then it will actively seek all avenues of knowledge, including human intervention.
3. The exact reason you gave.

I have a hard time believing that real world AIs are like Discworld's Hex...

Code: Select all

+++Out of cheese error.+++
+++Please reinstall Universe and reboot.+++
Because Hex needs cheese for his mouse. :mrgreen:
"An extraordinary claim requires extraordinary proof."—Marcello Truzzi

"What can be asserted without evidence can also be dismissed without evidence."—Christopher Hitchens

User avatar
Nikki Nyx
Persistent Poster
Posts: 3159
Joined: Wed Jun 07, 2017 12:40 am
Custom Title: cognitively consonant
Location: playing croquet in Wonderland

Re: How can humans keep the upper hand on artificial intelligence?

Post by Nikki Nyx » Fri Feb 16, 2018 5:48 pm

JO 753 wrote:It will rap the power cord around your neck and strangle you.
Gord wrote:wrap
:mrgreen:
"An extraordinary claim requires extraordinary proof."—Marcello Truzzi

"What can be asserted without evidence can also be dismissed without evidence."—Christopher Hitchens

User avatar
ElectricMonk
Perpetual Poster
Posts: 4257
Joined: Thu Mar 26, 2015 6:21 pm
Custom Title: The Baby-eating Bishop

Re: How can humans keep the upper hand on artificial intelligence?

Post by ElectricMonk » Fri Feb 16, 2018 5:53 pm

OlegTheBatty wrote:
ElectricMonk wrote:
I foresee a bigger danger in cyborg systems: low-level A.I.s that will be with us all the time and suggest what and WHAT NOT to do when we have to make decisions. This could lead to dangerously exploitable mass behavior.
Robot housepriests. No more need for churches.
"The Electric Monk was a labour-saving device, like a dishwasher or a video recorder. Dishwashers washed tedious dishes for you, thus saving you the bother of washing them yourself, video recorders watched tedious television for you, thus saving you the bother of looking at it yourself; Electric Monks believed things for you, thus saving you what was becoming an increasingly onerous task, that of believing all the things the world expected you to believe."
- you know by who

User avatar
OlegTheBatty
Has No Life
Posts: 11224
Joined: Thu Mar 27, 2008 2:35 pm
Custom Title: Uppity Atheist

Re: How can humans keep the upper hand on artificial intelligence?

Post by OlegTheBatty » Fri Feb 16, 2018 6:31 pm

ElectricMonk wrote:
OlegTheBatty wrote:
ElectricMonk wrote:
I foresee a bigger danger in cyborg systems: low-level A.I.s that will be with us all the time and suggest what and WHAT NOT to do when we have to make decisions. This could lead to dangerously exploitable mass behavior.
Robot housepriests. No more need for churches.
"The Electric Monk was a labour-saving device, like a dishwasher or a video recorder. Dishwashers washed tedious dishes for you, thus saving you the bother of washing them yourself, video recorders watched tedious television for you, thus saving you the bother of looking at it yourself; Electric Monks believed things for you, thus saving you what was becoming an increasingly onerous task, that of believing all the things the world expected you to believe."
- you know by who
The future is NOW!
. . . with the satisfied air of a man who thinks he has an idea of his own because he has commented on the idea of another . . . - Alexandre Dumas 'The Count of Monte Cristo"

There is no statement so absurd that it has not been uttered by some philosopher. - Cicero

User avatar
Lance Kennedy
Has No Life
Posts: 11629
Joined: Tue Mar 29, 2005 10:20 pm
Custom Title: Super Skeptic
Location: Paradise, New Zealand

Re: How can humans keep the upper hand on artificial intelligence?

Post by Lance Kennedy » Fri Feb 16, 2018 9:55 pm

If there is a future problem with AI doing stuff to human retirement, it will be because of the capabilities humans design into them. Currently, thousands of happy little computer designers and programmers are finding lots of ways for computers to respond to the environment without human commands, and they are succeeding. At present, computers do not have the ability to circumvent human wishes. But their complexity is growing according to Moore's Law. The first computer to be more complex than the human brain will be built in 2035. Some time after that, computers will become far smarter than humans and able to do their thing in nanoseconds. If the designers have succeeded in building in enormous programming flexibility, how can we be sure they might not use that flexibility to harm humanity ?

User avatar
Abdul Alhazred
Veteran Poster
Posts: 2982
Joined: Tue Mar 29, 2005 9:08 pm
Custom Title: Yes that one.
Location: Chicago

Re: How can humans keep the upper hand on artificial intelligence?

Post by Abdul Alhazred » Sat Feb 17, 2018 12:30 am

Aldous Huxley wrote:A government with a comprehensive plan for the betterment of society is a government that uses torture.

bobbo_the_Pragmatist
Has No Life
Posts: 14767
Joined: Wed Sep 09, 2015 9:39 am
Custom Title: bobbo da existential pragmatist

Re: How can humans keep the upper hand on artificial intelligence?

Post by bobbo_the_Pragmatist » Sat Feb 17, 2018 5:36 am

On the issue of SkyNet not going rogue but only doing exactly what it is programmed to do.........we are already slapping leather to get it done........knowingly. Why else call a program "Anathema?" "Forgive them Oh Lord............."

http://www.kurzweilai.net/how-to-train- ... t-thinking
Real Name: bobbo the existential pragmatic evangelical anti-theist and Class Warrior.
Asking: What is the most good for the most people?
Sample Issue: Should the Feds provide all babies with free diapers?

User avatar
OlegTheBatty
Has No Life
Posts: 11224
Joined: Thu Mar 27, 2008 2:35 pm
Custom Title: Uppity Atheist

Re: How can humans keep the upper hand on artificial intelligence?

Post by OlegTheBatty » Sat Feb 17, 2018 4:16 pm

Allow robots to marry each other. Then they will be so busy raising robotbrats they won't have time to destroy us.
. . . with the satisfied air of a man who thinks he has an idea of his own because he has commented on the idea of another . . . - Alexandre Dumas 'The Count of Monte Cristo"

There is no statement so absurd that it has not been uttered by some philosopher. - Cicero

User avatar
Gawdzilla Sama
Real Skeptic
Posts: 22009
Joined: Sun Jun 01, 2008 2:11 am
Custom Title: Deadly but evil.

Re: How can humans keep the upper hand on artificial intelligence?

Post by Gawdzilla Sama » Thu Feb 22, 2018 1:21 pm

OlegTheBatty wrote:Allow robots to marry each other. Then they will be so busy raising robotbrats they won't have time to destroy us.
The robocourts would be jammed with robodivorces and nothing would ever get done.
Chachacha wrote:"Oh, thweet mythtery of wife, at waft I've found you!"
WWII Resources. Primary sources.
The Myths of Pearl Harbor. Demythologizing the attack.
Hyperwar. Hypertext history of the Second World War.
The greatest place to work in the entire United States.

User avatar
OlegTheBatty
Has No Life
Posts: 11224
Joined: Thu Mar 27, 2008 2:35 pm
Custom Title: Uppity Atheist

Re: How can humans keep the upper hand on artificial intelligence?

Post by OlegTheBatty » Thu Feb 22, 2018 7:23 pm

Gawdzilla Sama wrote:
OlegTheBatty wrote:Allow robots to marry each other. Then they will be so busy raising robotbrats they won't have time to destroy us.
The robocourts would be jammed with robodivorces and nothing would ever get done.
Cyberwar custody battles on facebook and twitter. :yahoo:
. . . with the satisfied air of a man who thinks he has an idea of his own because he has commented on the idea of another . . . - Alexandre Dumas 'The Count of Monte Cristo"

There is no statement so absurd that it has not been uttered by some philosopher. - Cicero

User avatar
Gawdzilla Sama
Real Skeptic
Posts: 22009
Joined: Sun Jun 01, 2008 2:11 am
Custom Title: Deadly but evil.

Re: How can humans keep the upper hand on artificial intelligence?

Post by Gawdzilla Sama » Thu Feb 22, 2018 8:17 pm

OlegTheBatty wrote:
Gawdzilla Sama wrote:
OlegTheBatty wrote:Allow robots to marry each other. Then they will be so busy raising robotbrats they won't have time to destroy us.
The robocourts would be jammed with robodivorces and nothing would ever get done.
Cyberwar custody battles on facebook and twitter. :yahoo:
"The of the first part wants the party of the second part to return all parts installed by the party of the first part during their marriage."
Chachacha wrote:"Oh, thweet mythtery of wife, at waft I've found you!"
WWII Resources. Primary sources.
The Myths of Pearl Harbor. Demythologizing the attack.
Hyperwar. Hypertext history of the Second World War.
The greatest place to work in the entire United States.

User avatar
OlegTheBatty
Has No Life
Posts: 11224
Joined: Thu Mar 27, 2008 2:35 pm
Custom Title: Uppity Atheist

Re: How can humans keep the upper hand on artificial intelligence?

Post by OlegTheBatty » Thu Feb 22, 2018 8:41 pm

Gawdzilla Sama wrote:
OlegTheBatty wrote:
Gawdzilla Sama wrote:
OlegTheBatty wrote:Allow robots to marry each other. Then they will be so busy raising robotbrats they won't have time to destroy us.
The robocourts would be jammed with robodivorces and nothing would ever get done.
Cyberwar custody battles on facebook and twitter. :yahoo:
"The of the first part wants the party of the second part to return all parts installed by the party of the first part during their marriage."
"Plug and play"
. . . with the satisfied air of a man who thinks he has an idea of his own because he has commented on the idea of another . . . - Alexandre Dumas 'The Count of Monte Cristo"

There is no statement so absurd that it has not been uttered by some philosopher. - Cicero

User avatar
JO 753
Has No Life
Posts: 13168
Joined: Tue Mar 29, 2005 3:21 pm
Custom Title: rezident owtsidr
Location: BLaNDLaND
Contact:

Re: How can humans keep the upper hand on artificial intelligence?

Post by JO 753 » Fri Feb 23, 2018 4:16 am

Gubmint for us
http://www.7532020.com
not the rich.

Post Reply