Scientists Worry Over Super-Smart AI

By Kevin Parrish, published on July 27, 2009 at 4:10 PM
Source: Tom's Guide US | Keywords: , , , , | Themes: Business
Syndication: Add to your Google homepage Add to My Yahoo!

Will 2001: A Space Odyssey's HAL 9000 become a reality soon? No, but scientists fear that technology is heading that way.

This weekend John Markoff of The New York Times wrote an interesting article about machines, and how they may eventually outsmart man. His opening paragraph describes three scenarios that are already a reality: a robot that can open doors and track down an electrical outlet to recharge itself, machines that are very close to killing humans autonomously, and unstoppable computer viruses and worms that have reached a "cockroach" stage of machine intelligence.

The good news is that artificial intelligence hasn't reach the "HAL 9000" level of intellect; computers haven't become self-aware, nor will they form any kind of Skynet any time soon. However many researchers have agreed that the killing robots, as previously mentioned, are in fact here, or will be here soon.

Additionally, progress has raised concerns that robots will take the place of human workers, and that humans will eventually be forced to live with machines that mimic human behaviors. There's also concern that criminals could take advantage of advancements in AI, using a "speech synthesis system" to impersonate another human, for example, or mining smart phones to uncover personal information.

Does that mean super-intelligent machines and artificial intelligence systems will eventually run amok? The researchers attending the conference (mostly) discounted the idea of a spontaneous intelligence stemming from the Internet, and other highly centralized intelligences outside the web. However, Dr. Horvitz said that computer scientists must respond to the notions nonetheless.

To read a more detailed version of the article, pick up a weekend copy of the New York Times or head here.

Comments | Print | Send to a friend

Sponsored links

Comments

redkachina 07/27/2009 10:16 PM
Hide
-3+

There's always Isaac Asimov's 3 Laws of Robotic..LoL

lifelesspoet 07/27/2009 10:22 PM
Hide
-2+

If Asimov's three rules don't work, there is another way.
Here are my three rules for preventing a robot Apocalypse, or at least to prevent another atrocity that was the I, robot movie.
Not allowing freedom of movement.
No direct connection to the internet or other AI's.
Hard wired failsafes and manual back up systems.

DjEaZy 07/27/2009 10:36 PM
Hide
-2+

... it's a fascinated time to be alive... all, that we need, is a shift from monetary based to resource based economy... because when not, then "Additionally, progress has raised concerns that robots will take the place of human workers, and that humans will eventually be forced to live with machines that mimic human behaviors. There's also concern that criminals could take advantage of advancements in AI, using a "speech synthesis system" to impersonate another human, for example, or mining smart phones to uncover personal information.
"

tipoo 07/27/2009 10:45 PM
Hide
-0+

Wasnt IBM trying to make a supercomputer with the intelligence of a dog or something like that? What ever happened with that?

syavash 07/27/2009 10:46 PM
Hide
-2+

Well if you take this think in the right direction

just think, robots doing all the work for free, no need to pay them, MEN WILL BE GREE TO FOLLOW THEIR PLEASURES AND spend mor etime inventing then acting like a workhorse

ckthecerealkiller 07/27/2009 10:53 PM
Hide
-5+

Oh noes we are all going to die! Where's me flyin cars already!?!?

daedalus685 07/27/2009 10:57 PM
Hide
-0+

"Will 2001: A Space Odyssey's HAL 9000 become a reality soon? No, but scientists fear that technology is heading that way."

Human level intelligence in AI is far closer than most people think, or want to accept. Given exponential growth of most information technology it isn't unreasonable to excpect 'HAL' level machines in the next 15 years..

turboflame 07/27/2009 11:00 PM
Hide
-7+

I for one welcome our new robot overlords.

jhansonxi 07/27/2009 11:06 PM
Hide
-20+

Rogue AIs recommend Windows Vista® Home Premium

AMDnoob 07/27/2009 11:14 PM
Hide
-11+

just install Norton on every computer in the world, there problem solved! lol

Upendra09 07/27/2009 11:23 PM
Hide
-12+

But can they run Crysis? LOL

Any way robots will be cause of the next great depression because the will put tons of humans out of work

will_chellam 07/27/2009 11:29 PM
Hide
-1+

The robots that autonomously kill, I seem to remember they were made by samsung - the face recognition on their digicams seems ok-ish, so im sure a robot mounted machine gun will be fine.....

marnad 07/27/2009 11:46 PM
Hide
-1+

Hmm, quite interesting. I'm sure, as others probably are, that many of the critical advancements in making robots with human-like intelligence has occured in the private sector of development, as opposed to the public eye.

I'm betting there is a lot of concern behind closed doors. Allow yourself to envision robots that can establish their own organizations, mine for elements, improve themselves, etc. It's a sobering thought. Some action should be taken to prevent selecting the wrong path with robotics development.

crom 07/27/2009 11:47 PM
Hide
-10+

As long as Windows is the prevalent operating system I don't think we have anything to fear from the robots... until they get SP2!

pirateboy 07/27/2009 11:55 PM
Hide
-3+

I bet most of you girlfriendless nerds would love to own a cherry2000 and raise her in your mom's basement ;)

Ciuy 07/28/2009 12:14 PM
Hide
-1+

and little green men will come to rescue us from the robots .

Balshoy 07/28/2009 12:37 PM
Hide
-0+

I for one welcome my cyborg masters! lol :D

Raid3r 07/28/2009 12:39 PM
Hide
-0+

Such a slippery slope..for not much gain.

hercules 07/28/2009 12:52 PM
Hide
-2+

Bring on Ghost in the shell!

duesouth 07/28/2009 12:54 PM
Show
Cache 07/28/2009 1:11 AM
Hide
-4+

I always find it amusing that the concept of a true breakout AI somehow will always turn against the humans that created it. There is always the chance that AI will ultimately look at us with the eyerolls we gave our parents when we were teens. Ultimately, anything that is 'alive' in a technical sense must want something. As human beings, we want food, love, shelter, companionship, the ability to mate, to raise children, etc. We make things happen precisely because we need things.

What would a machine want, and what could it really do to achieve those goals? Without mobility and the capability to achieve complex tasks--the first of which would have to be securing power so that it would not die--it would simply wither and eventually die out, incapable of realizing any goal save making its existence known.

ak47lt 07/28/2009 1:17 AM
Hide
-0+

The most complex Neural Networks (it is the only option to really emulate complex decision making) using the most powerful computers today can emulate the intellect of the bee. New operating systems or better hardware do not have much effect here - better algorithms are needed.
Some of us has masters degree in computer science :)

djcoolmasterx 07/28/2009 1:22 AM
Hide
-1+

Actually, that version sucked. This one: http://www.youtube.com/watch?v=WGoi1MSGu64

matt87_50 07/28/2009 2:05 AM
Hide
-0+

and I assume there are no photos of the event, as they were afraid of having their souls stolen

marraco 07/28/2009 2:20 AM
Hide
-3+

syavash :
Well if you take this think in the right directionjust think, robots doing all the work for free, no need to pay them, MEN WILL BE GREE TO FOLLOW THEIR PLEASURES AND spend mor etime inventing then acting like a workhorse


well. my computer takes a fair amount of my wage...

I guess my robot will annoy me continously about his obsolete memory, procesors, OS, and wathever... :)

Greg_77 07/28/2009 2:49 AM
Hide
-1+

Anyone for a robot overclocking contest?

doomtomb 07/28/2009 3:12 AM
Hide
-0+

Military is the source of this scary AI. Soon we'll see stuff out of the movie Stealth.

rambo117 07/28/2009 3:21 AM
Hide
-1+

well, on december 21st 2012, looks like skynet will become self aware

gamefreak62 07/28/2009 4:06 AM
Hide
-5+

Ridiculous. Computers do what we program them to do, they don't spontaneously do things. People program them that way. If someone programs them to be malicious, then they will be malicious. Besides, the way computers are breaking these days we'd only have to hide for about ten years for all the computers to blue screen and need to be rebooted.

rambo117 07/28/2009 4:08 AM
Hide
-1+

^1+

Computers are not human, and never will be.
Unless we build an AI that adapts, but that still wouldnt be human.

like all other machines, they have an off switch


Sponsored links