Board index › Topical Discussion › Politics, Philosophy, and Religion

Humanity obsolete?

Page 1 of 2 [ 23 posts ]  Go to page 1, 2  Next

Previous topic | Next topic

Sand
Veteran
Veteran

User avatar

Joined: 15 Sep 2007
Age: 100
Gender: Male
Posts: 11,484
Location: Finland

10 Feb 2010, 8:12 pm

See http://hplusmagazine.com/articles/ai/ho ... n-level-ai



Sand
Veteran
Veteran

User avatar

Joined: 15 Sep 2007
Age: 100
Gender: Male
Posts: 11,484
Location: Finland

11 Feb 2010, 9:13 am

No comments so far. Not much interest in humanity creating its own aliens to take over. This item from Slashdot might bring it a little closer.

"British criminals should soon prepare to be shot at from unmanned airborne police robots. Last month it was revealed that modified military aircraft drones will carry out surveillance on everyone from British protesters and antisocial motorists to fly-tippers. But these drones could be armed with tasers, non-lethal projectiles and ultra-powerful disorienting strobe lighting apparatus, reports Wired. The flying robot fleet will range from miniature tactical craft such as the miniature AirRobot being tested by one police force, to BAE System's new 12m-wide armed HERTI drone as flown in Afghanistan."

When the AI gets control of these cute items humanity might reconsider the wisdom of moves in this direction.



PLA
Veteran
Veteran

User avatar

Joined: 10 May 2007
Age: 36
Gender: Male
Posts: 1,929
Location: Sweden

11 Feb 2010, 10:31 am

It'll be fine if we can find some little runt with a knack for making Arnold Schwarzenegger do his bidding.


_________________
I can make a statement true by placing it first in this signature.

"Everyone loves the dolphin. A bitter shark - emerging from it's cold depths - doesn't stand a chance." This is hyperbol.

"Run, Jump, Fall, Limp off, Try Harder."


Orwell
Veteran
Veteran

User avatar

Joined: 8 Aug 2007
Age: 36
Gender: Male
Posts: 12,518
Location: Room 101

11 Feb 2010, 11:41 am

I find it interesting that some people claim AI will be doing Nobel-quality scientific work before they can pass the Turing test, on the basis that a Turing AI would have to conceal some of its intelligence and feign human foibles. If this is true, I doubt whether passing the Turing Test is a valuable goal or just a gimmick, and I would also question the wisdom of making such an AI.


_________________
WAR IS PEACE
FREEDOM IS SLAVERY
IGNORANCE IS STRENGTH


PLA
Veteran
Veteran

User avatar

Joined: 10 May 2007
Age: 36
Gender: Male
Posts: 1,929
Location: Sweden

12 Feb 2010, 3:12 am

Well, the Turing test isn't supposed to be the only measure of intelligence, rather a sufficient measure of indistinguishability.


_________________
I can make a statement true by placing it first in this signature.

"Everyone loves the dolphin. A bitter shark - emerging from it's cold depths - doesn't stand a chance." This is hyperbol.

"Run, Jump, Fall, Limp off, Try Harder."


PlatedDrake
Veteran
Veteran

User avatar

Joined: 25 Aug 2009
Age: 45
Gender: Male
Posts: 1,365
Location: Piedmont Region, NC, USA

12 Feb 2010, 9:09 am

I wont mind AI so long as the "Three Laws" are in place . . . or, it'll have the mindset of an autist (meaning that it has no need to harm others just because it wouldnt make any sense, would be emotionally detached and not subject to bias, and would likely be addicted to information).



mjs82
Veteran
Veteran

User avatar

Joined: 20 Jun 2005
Age: 44
Gender: Female
Posts: 3,166

12 Feb 2010, 9:28 am

If AI's take over, then we were stupid enough to let it happen. If they don't take over, we were stupid enough to not let it happen.



zer0netgain
Veteran
Veteran

User avatar

Joined: 2 Mar 2009
Age: 58
Gender: Male
Posts: 6,615

12 Feb 2010, 9:35 am

Orwell wrote:
I find it interesting that some people claim AI will be doing Nobel-quality scientific work before they can pass the Turing test, on the basis that a Turing AI would have to conceal some of its intelligence and feign human foibles. If this is true, I doubt whether passing the Turing Test is a valuable goal or just a gimmick, and I would also question the wisdom of making such an AI.


Enlighten me, please.

As far as I know, a real "AI" would have to have the ability to learn, grow and adapt without aid from a programmer. What does the Turing test require for something to be considered an AI?



Tensu
Veteran
Veteran

User avatar

Joined: 30 Dec 2009
Age: 37
Gender: Male
Posts: 1,661
Location: Nixa, MO, USA

12 Feb 2010, 9:46 am

So long as we do not foolishly make any robots magnet-proof, we will hold the edge over any robot we build, no matter how advanced it's AI.



Orwell
Veteran
Veteran

User avatar

Joined: 8 Aug 2007
Age: 36
Gender: Male
Posts: 12,518
Location: Room 101

12 Feb 2010, 9:52 am

zer0netgain wrote:
Orwell wrote:
I find it interesting that some people claim AI will be doing Nobel-quality scientific work before they can pass the Turing test, on the basis that a Turing AI would have to conceal some of its intelligence and feign human foibles. If this is true, I doubt whether passing the Turing Test is a valuable goal or just a gimmick, and I would also question the wisdom of making such an AI.


Enlighten me, please.

As far as I know, a real "AI" would have to have the ability to learn, grow and adapt without aid from a programmer. What does the Turing test require for something to be considered an AI?

The Turing Test is a measure of how closely an AI approximates a human. The idea is that when a human can not tell whether they are conversing with another human or with the AI, it has passed the Turing Test. According to some people in the OP's article, to pass such a test an AI would have to feign emotions, irrationality, etc, as well as hide its superior intelligence (presuming it had such).


_________________
WAR IS PEACE
FREEDOM IS SLAVERY
IGNORANCE IS STRENGTH


Sand
Veteran
Veteran

User avatar

Joined: 15 Sep 2007
Age: 100
Gender: Male
Posts: 11,484
Location: Finland

12 Feb 2010, 10:00 am

Once you create something more intelligent than yourself you've got a wild cannon on your hands. You cannot control it.



PLA
Veteran
Veteran

User avatar

Joined: 10 May 2007
Age: 36
Gender: Male
Posts: 1,929
Location: Sweden

12 Feb 2010, 10:28 am

I disagree. Less intelligent things have exerted control over more intelligent things before.

Arguably. It depends on what we are referring to as "intelligence" here.


_________________
I can make a statement true by placing it first in this signature.

"Everyone loves the dolphin. A bitter shark - emerging from it's cold depths - doesn't stand a chance." This is hyperbol.

"Run, Jump, Fall, Limp off, Try Harder."


mgran
Veteran
Veteran

User avatar

Joined: 22 May 2009
Gender: Female
Posts: 1,864

12 Feb 2010, 10:35 am

Sand wrote:
No comments so far. Not much interest in humanity creating its own aliens to take over. This item from Slashdot might bring it a little closer.

"British criminals should soon prepare to be shot at from unmanned airborne police robots. Last month it was revealed that modified military aircraft drones will carry out surveillance on everyone from British protesters and antisocial motorists to fly-tippers. But these drones could be armed with tasers, non-lethal projectiles and ultra-powerful disorienting strobe lighting apparatus, reports Wired. The flying robot fleet will range from miniature tactical craft such as the miniature AirRobot being tested by one police force, to BAE System's new 12m-wide armed HERTI drone as flown in Afghanistan."

When the AI gets control of these cute items humanity might reconsider the wisdom of moves in this direction.
Hey Sand, could you pm me a link to research this? I'm very interested.



Orwell
Veteran
Veteran

User avatar

Joined: 8 Aug 2007
Age: 36
Gender: Male
Posts: 12,518
Location: Room 101

12 Feb 2010, 10:46 am

Sand wrote:
Once you create something more intelligent than yourself you've got a wild cannon on your hands. You cannot control it.

Hopefully the creators will have the common sense to build in safety mechanisms in places the AI can not access, in order to prevent it from disabling said safeties. I would be wary of giving an AI access to networks.


_________________
WAR IS PEACE
FREEDOM IS SLAVERY
IGNORANCE IS STRENGTH


Sand
Veteran
Veteran

User avatar

Joined: 15 Sep 2007
Age: 100
Gender: Male
Posts: 11,484
Location: Finland

12 Feb 2010, 11:13 am

mgran wrote:
Sand wrote:
No comments so far. Not much interest in humanity creating its own aliens to take over. This item from Slashdot might bring it a little closer.

"British criminals should soon prepare to be shot at from unmanned airborne police robots. Last month it was revealed that modified military aircraft drones will carry out surveillance on everyone from British protesters and antisocial motorists to fly-tippers. But these drones could be armed with tasers, non-lethal projectiles and ultra-powerful disorienting strobe lighting apparatus, reports Wired. The flying robot fleet will range from miniature tactical craft such as the miniature AirRobot being tested by one police force, to BAE System's new 12m-wide armed HERTI drone as flown in Afghanistan."

When the AI gets control of these cute items humanity might reconsider the wisdom of moves in this direction.
Hey Sand, could you pm me a link to research this? I'm very interested.


See
http://www.wired.co.uk/news/archive/201 ... rones.aspx



Sand
Veteran
Veteran

User avatar

Joined: 15 Sep 2007
Age: 100
Gender: Male
Posts: 11,484
Location: Finland

12 Feb 2010, 11:19 am

Orwell wrote:
Sand wrote:
Once you create something more intelligent than yourself you've got a wild cannon on your hands. You cannot control it.

Hopefully the creators will have the common sense to build in safety mechanisms in places the AI can not access, in order to prevent it from disabling said safeties. I would be wary of giving an AI access to networks.


There is no guarantee a super intelligent AI can be out thought. Whatever safeguards are set can probably be defeated. "Blade Runner" indicated a limited life might work but I doubt it.