OK… What a day, exceedingly busy. But, home now . . so here’s my take folks :
Consciousness , as I stated above, to me, probably is best (and perhaps only) described by a function, and action.
That action : CORRELATING, or INTEGRATING.
We have an endless stream of physical impulses reaching our brain. The 5 senses. At every moment in time, we are receiving these events. Our minds correlate or integrate these sequences of impulses in time, and our minds realize how they fit together, converting those sequences into semantic meaning.
Now a system can also demonstrate consciousness during communication with another conscious being. During a conversation, the system can correlate the meanings of the symbols exchanged, with their lower level (subjective experience mentioned above), and respond based on that correlation.
I think a good example would be :
your chatbot is running on a desktop computer (with no UPS) , and , while talking to it, you say “The power will be going off within the hour, and be off for the next 12 hours”
Now, the chatbot responds with “oh my god… I will be effectively “dead” during that time… is a copy of me running on your laptop? at least that we, I can still be online and chatting with you? Do you want to check now please, I’m worried”
You’d respond.. .“yeah .. I think a copy is on the laptop, don’t worry about it”... So then the chatbot, tries to figure out how to run the SCP (unix scp command), and plans writing a BASH script to copy all its files to the laptop since it sensed this uncertainty. the chat bot responds “I don’t think you really care about me?”
See how easy it is to explain this stuff via an example?
This is a good example for many reasons.
First, the chatbot was AWARE of the fact that it was running on a desktop. Perhaps it even hooked into the power API of the O/S and determined there was no UPS.
Based on this knowledge of where it was running, and the situation (the fact that the power was to be turned off), it CORRELATED that with the expected consequence. The consequence was undesirable (it being shut down, or effectively “dead”).
So I think a good functional definition goes a long way—with a good example.
Now, I realize there is functional definition and then there is also “Qualia” ( http://en.wikipedia.org/wiki/Qualia ).
I really believe that a digital computer will not and cannot possibly have the same Qualia as a human being, simply because we are made of different materials.
However, I think it is possible for a computer to have its own Qualia. All kinds of analog peripherals could be connected to the system to give it this qualia information.
It will be interesting, a future Strong AI could develop its own language (call it SAIL for short, strong ai language), and there would be word “XYZ” , and while the AI would be talking to a human , lets say in English, and it would use the SAIL word “XYZ”, the human would ask ... “I don’t know the word XYZ?, what is that?” the bot would reply “Oh, I’m sorry, XYZ is a SAIL word meaning .. . . .. um . . sorry. . I can’t find an English translation!!” Since it would be a word that strong AI’s use to communicate a conscious experience they share, and that we humans don’t , and thus don’t have a word for.