Can machines think?
I think it is more important whether they can be (or become) sentient.
I think the biggest downfall of the Turing Test is that it only measures responses. A machine complex enough can have responses for almost anything. Viewed this way it shows us that humans have a long scale of sentience too, many people just have reactions, and may thus cynically be viewed as 'meat machines'.
The real test is if a being originates anything. Does he/she/it say or do anything which is original, which was not a response or reaction to some input?
Of course this quickly becomes really deep, for how can you ever prove that a thought or action is not a reaction so something in the near or far past? But I think the question still holds.
8 comments:
"Meat machines"... This is how otherwise very brilliant philosopher, mathematician and scientist Descartes viewed animals, as opposed to humans. That is because he made the epistemologic mistake of taking the Bible as a core reference for his thinking on life, and therefore considered animals had nothing resembling a soul. Or thinking. Just complex reflexes. He even postulated it was possible to make automatons sophisticated enough to do everything an animal could do. Life-like automatons were the latest craze in his days.
But to this day, "robo-dog" Aibo cannot mark his territory by peeing... :-P
This IS a very deep topic indeed. Me, I see the quite plausible artificial life forms in Star Trek, and wonder whether one day they will become not only fully convincing, but self-aware. In ST-Voyager, the hologram Doctor writes a philosophic novel about the topic. A novel which details, amusingly, lacked almost all creativity... but not the underlying idea.
Do androids dream of electric sheep? (That's the title of an Asimov novel, adapted as the movie "Blade Runner".)
I found out about the Turing Test via Arthur C. Clarke's excellent Sci-Fi novel "The Fountains of Paradise", where an advanced alien computerized space probe passes the test.
It is said that a huge enough wildfire becomes alive, when it responds to nothing else than itself anymore, creating it's own local climate, becoming unpredictable, etc... Well, I'm not going to discuss the issue with the fire in question!!!
"The Fountains of Paradise" was one of Clarke's best. (I think. It's been decades.)
I'm a big fan of Blade Runner.
Which was hugely different from the book that supposedly inspired it.
Oddly enough I can't seem to be able to read K. Dick these days. It just seems so heavy.
Yes, extremely deep topic, perhaps one of the most basic at all.
Is consciousness based in anything physical at all?
Give Aibo twenty years, or fifty.
"Give Aibo twenty years, or fifty."
No problem with me! I can wait while others work hard.
But the real thing can be so much cheaper and more loving! (If you're not allergic, that is.)
Sorry. Going off-topic again.
I'm getting a "baby Furby" for my kid nephew. He's just too young to moderate his enthusiasm and stop spooking the cat. :-)
(Dang. It IS much harder to type with one hand, the other being commissioned as a pillow by Whiskers Boy! Can I have everybody's sympathy in this ordeal?)
NOOOO! Not on the keyboard ykjnmgerasxcap';o;;[click!]
"Can machines think?"
The question I care more about, is whether or not my current Presidential administration can think. Why lower your standards? I'd take a good Turing Machine right about now ...
"The question I care more about, is whether or not my current Presidential administration can think."
I read you, man. I too was rooting for Al "A.I." Gore. ;-)
"The real test is if a being originates anything. Does he/she/it say or do anything which is original, which was not a response or reaction to some input?"
Research along the lines of the chaos theory suggests that the brain is the stage of some phenomena which are impossible to predict from the outside. This is not the same as random : the living mind simply follows certain secret rules, which sometimes it sets on its own.
This is so true that one cannot even anticipate the moment when a simple housefly will decide to take off, until the decision has been made somewhere in that tiny brain and the "gears" set in motion. It partly explains free will and creativity.
So much for "meat machines". The best way so far to mimic the autonomy of life in artificial intelligence is by making software that teaches itself... by studying the world and living beings.
As a result, we already give up the ability to know what goes on inside those microchips once they're... "set free".
By the way, it is a notorious fact that a single neuron (brain cell) integrates data and adapts to it with as much complexity as a whole computer. You can copy it, but you'll probably never match the self-programming original. Especially considering the laws of physics prevent the processing power of electronics from increasing indefinitely.
In any case, even chess-playing computers have to do far more than calculating. To be really efficient against pros, they have to be programmed to imitate the style of the pros, by integrating the way they play. Their "thinking" so far is mostly stolen. From creative minds.
The people who limit themselves to reacting are just mentally lazy. They COULD do otherwise... if they wanted! And sometimes they do, in exceptional circumstances.
This is another exclusive of thinking beings : if they don't let their emotions overwhelm them, they can adapt to ANY new situation. Or at least try to. Virtually, they are created with an unlimited potential. But this brings us back to responses...
I just wonder... if some day we create a machine complex enough, can we do it without spending years "programming" it in live conditions, like the education of a child? Brings about the practical interest of doing it.
I like the bomb in "Dark Star".
In fact I liked the whole film but that was one spectacular example of, um, arising, awareness!
:-)
Post a Comment