The phenomenology of the Turing test
Can machines think? Not yet, however soon they will. Place a powerful
computer (network) with the Google knowledge base behind a screen,
interrogate it and you won’t be able to distinguish its response from
that of a human. By the way, you interrogate the computer using text
formatted linguistic input. This is the essence of the Turing test.
With Google at hand, soon it will be even smarter than you.
Turing believed that appropriately programmed computer can think. Or
better the computer simulates thinking which cannot be distinguished
from your thinking. The digital computer does not posses intelligence,
it simulates it. Some believe that soon it will simulate a mind or
even consciousness. All you need is a powerful computer, and since computer
hardware develops exponentially (Moore’s law), soon conscious computers
(agents, robots) will hop around.
This wishful thinking illustrates yet another shortcoming of AI (Artificial
Intelligence) reasoning, Anthropomorphism. The other,
“Cartesian Slumber”,
was discussed in a previous chapter.
Following the Creator who created us in His image, AI seeks to create
machines in the human image. The only meaningful intelligence
is human. Which is the rational behind the Turing’s test, which tests
a dis-embodied intelligence. What about a Turing test which allows
us to talk to the computer? Since voice reveals emotions, we might
easily spot the culprit (simulator). Will the future computer simulate
emotions? Since emotions require embodiment, the question has to be
rephrased, how to create an embodied zombie which will fake (simulate)
emotions? Yet another anthropomorphism.
Is swarm intelligence a real intelligence or a misnomer? Does an ameba
possess an intelligence despite not being able to convey it to us?
AI would dismiss such speculations as nonsense. Let’s turn therefore
to Hans Jonas’ “The phenomenon of Life” (1): Plants,
animals and the human animal display an ascending development
of organic functions and capabilities. The emergence of the
human mind does not mark a great divide within nature but elaborates
what is prefigured throughout the life-world. The organic even
in its lowest forms prefigures mind, and the mind even on its
highest reaches remains part of the organic.
In other words, the organic even in its lowest forms prefigures intelligence,
which shapes also the human intelligence. AI ought to study first
the intelligence of an ameba. Simulate it and model it. This basic model
may then serve as an initial state from which intelligence, and even
consciousness might emerge. Such an endeavor requires a new kind of
programming tool different from what is known today. It might look like
the one which I apply in my CA studies. You plant two zygotes which
evolve into a stable system with emerging
and unpredictable properties.
Additional reading Robot mind
References
1. Hans Jonas The Phenomenon of Life- Toward a Philosophical Biology
Northwestern University Press Evanston Ill 2001
Back to complexity index