Search This Blog

A Better Turing Test

What is consciousness?
What will it take for computers to learn to think? Do we have to understand consciousness to build a machine that has it? How does the brain do what it does? How can we find out? What will happen after the moment consciousness rises from machinery? Some people, my self included, believe the Singularity is not unlike the event horizon at the edge of a black hole, a point of no return beyond which acceleration towards complexity so profoundly outstrips the ability of biology or human culture as to negate our capacity for perception... we simply will not be able to see it!

The IEEE has compiled a special section (The Singularity: A Special Report, June, 08) on the issues and implications of what AI, Science Fiction, and Futurists call the "Singularity"... a moment when machines reach the ability to consciously close their own self evolutionary loop... physical evolution at the pace of knowledge or representation...

What do you think?

Funny that there are such huge differences when comparing the requirements of a Turing Machine (minimally complete universal algorithm cruncher) and the Turing Test (minimally conscious intelligence). There are some who believe a Turing complete Machine (and the right software) is all you need to satisfy a Turing Test for machine intelligence. And, there are others who say that intelligence can emerge only from machines of far greater structural complexity. 

My own personal view is that the only difference is speed. If you don't mind waiting (a long long time), a Turing complete Machine executing the correct instructions, should be able to pass any Turing Test. The corollary is that the human mind, no matter how complex, is fundamentally just a Turing machine. Structural complexity can provide quantitative but never qualitative advantage.  This argument will be substantially influenced by a greater understanding of quantum physics.  If the mind is a quantum machine and if quantum physics defines information differently, then all bets are off.  So far I see zero evidence that biology works by quantum rules or has evolved the chemical or mechanical armatures necessary to reach down and translate quantum causality up and out to our classical relativistic world.  There is so much confusion about this.  Of course everything in the classical world is quantum derived, but quantum derived structure and behavior played out in large numbers is exactly why the classical world is so... classical.  Interacting with the quantum world means isolating individual quanta.  No trivial task.

Cultural bias, anthropomorphism, human centered perspective weighs heavily on our ability to clearly reason about this issue. There is something comforting in imagining the human mind as special, as different, as more than mechanical, as inimitable, outside of the confines of physics. It is much harder to imagine a machine who's mechanical functioning results causally in the sensations and experiences we call mind, emotions, consciousness, and soul.