Science

Sony Expert Says Turing Had it Wrong

Posted 17 Aug 2002 at 05:14 UTC by steve Share This

According to Luc Steels, director of the Sony Computer Science Lab, Alan Turing had it all wrong when he proposed that a successful artificial intelligence should be able to converse like a human. Steels believes robots will develop their own type of intelligence through interaction with their own kind and their own culture. For more details read the CNet story.


Can't get away from the flapping wings, posted 17 Aug 2002 at 05:19 UTC by steve » (Master)

I just had to post this story in light of the controversy of the last one. Not only because he mentions the "E" word a few times but because of this ironic quote:

"It's like saying you want to make a flying machine, so you produce something that is indistinguishable from a bird. On the other hand, an airplane achieves flight but it doesn't need to flap its wings."

"E word", posted 17 Aug 2002 at 21:31 UTC by robodave » (Journeyer)

You forgot to mention Evolution Robotics with their image recognizing and navigating robot, not dissimilar in construction from Rog-o-matics computer chair.

What kind of intelligence?, posted 18 Aug 2002 at 05:29 UTC by jeffkoenig » (Master)

Unfortunately for Mr. Steels (and maybe ourselves), the only metric we presently have for gauging intelligence is our own.

Turing said that machines would attain "intelligence" when they would become indistinguishable from a human in conversation. I think this was his 29th argument out of 150, and has come to be called the "Turing Test".

Yes, perhaps some new kind of intelligence could <cough> "evolve". But, 1) How would we know, and 2) Has this happened, but we are incapable of understanding it?

Maybe computers will someday conjure up a "Steels Test" to see if humans are intelligent...

What robots should really do, posted 18 Aug 2002 at 15:44 UTC by ROB.T. » (Master)

Honestly people, if I wanted to build robots that formed their own "culture" and developed their own "minds" I would simply have more kids.

What I want is a robot that will do exactly what I tell it to do in just the way I intended it. When I say "mow the lawn" I expect the whole lawn to be mowed front AND back neat and proper. No discussion, no "I want to do it this way", it just gets done and my 'bot tucks itself safely away without demanding a higher allowance.

Sometimes I wonder what color the sky is the world of some of these college professors.

ROB.T.

Robot Explorations, posted 18 Aug 2002 at 22:24 UTC by robodave » (Journeyer)

I'm not sure what color sky they do see, but perhaps this is a next step.

Robots at present can be programmed and built to accomplish specific tasks, from moving a side panel of a car from a feeder to an assembly, to mowing a lawn.

But what about an unspecific task? Mowing your neighbors lawn? Mowing without touching the flower bed? Crossing the street to get to the neighbors? Knowing that it's time to mow the lawn with respect to grass height and the sprinklers aren't running and it isn't raining and the noise won't wake the neighbors and no one is playing on the lawn? Avoiding the misthrown newspaper? These are all pretty specific and could be programmed for.. Hedge trimming.. Noticing pests.. edging.. lawn health.. Cleaning the pool.. checking pool water.. getting me an ice tea.. the tasks begin to multiply.. Hold a conversation, there's the Turing part, a start of engaging a person, entertainment as sought by Sony with the Aibo and SDR-4X robots. Is conversation, entertainment a specific task? Can a robot be specifically programmed to maintain human interest, to inform with new information, to enrich a humans social life? If you say that this task is un-needed, when was the last time you (or I) visited a retirement home? Talked with your kids? Do we need robots to help us with the caring we should be doing? Is this where a generalized robot starts to come in, one that might query as to specific needs and accomplish them "just the way I intended it"?

Ya know, looking back at this and rereading it, it does sound pretty weird. Just some different ideas on what "robot" can mean to different folks. People generally can't define what specifically a robot is, hence we have BEAM robots, Battlebots, and discussions as to whether our microwave is a robot.

Talking to the Doll Fans, posted 18 Aug 2002 at 22:27 UTC by The Swirling Brain » (Master)

If robots could interact and develop their own intelligence, then... 1. We'd have to figure out how to translate their communications. (What's that movie about Guardian and what was that other computer's name?) 2. Phlisophers would have to spend eons deciding if the computers really were thinking or not and if they were alive or not. 3. We'd wonder what they think of us. 4. We might finally stop calling it "artifical" intelligence. 5. They'd decide our fate in a microsecond, they'd build skynet and we'd soon be replaced by them. 6. The dalecs would then exterminate all life. 7. If that were to happen would God hit the reset button. 8. We'd go through eons of chkdisk. 9. We'd reboot into Windows. 10. Ultimately that would bring us to the blue screen of death.

Deep Thought's by Douglas Adams, posted 18 Aug 2002 at 22:31 UTC by The Swirling Brain » (Master)

Of course, you know the answer is 42.

von Neumann vs. Alan Turing, posted 19 Aug 2002 at 20:03 UTC by evilrobots » (Observer)

von Neumann is the father of all robos, Alan Turing is the mother of all beautiful minds. that's my 2 cents.

See more of the latest robot news!

Recent blogs

26 Nov 2014 shimniok (Journeyer)
15 Nov 2014 mwaibel (Master)
14 Nov 2014 Sergey Popov (Apprentice)
14 Nov 2014 wedesoft (Master)
5 Aug 2014 svo (Master)
20 Jul 2014 Flanneltron (Journeyer)
3 Jul 2014 jmhenry (Journeyer)
3 Jul 2014 steve (Master)
2 Jul 2014 Petar.Kormushev (Master)
10 Jun 2014 robotvibes (Master)
X
Share this page