I started this business law class as well as a Science and Technology class that sought to teach the history of technology, and I was thinking about my research on AI and classes that teach about ancient societies...
I think we've already created ai 1000's of years ago with forms of governments and societies.
Our laws and industries make up the "code" and "functions" of society.
Our "research"/technology is the product of our "ai".
Our legislative body and makeup of statute's and delegation to agencies, creates basically "self-organizing" code/sub-functions.
We can compare statutes and administrative bodies changes of legal conduct with diff patches...
Yeah, deep right?
The coming together [of ourselves as a society] shows that the whole is greater than the sum of the parts. By combining, we can create more intricate ideas and technologies.
It's weird, but it's almost as if societies have a soul, a zeitgeist if you will.
The only catch:
Societies can't operate without a recollection of past events.
Which requires individuals remembering things, but soon databases will do that...
Update. Thanks virmaior.
Aristotle had hit on this concept early in Greek philosophy. I think Aristotle held the Greek state as superior to all the other states for it's purity, but that is conjecture. It may be why he advised Alexander to besiege his enemies as if they were beasts but his kinsman as equal or something. I think I presented this question because it blew my mind that societies like Ancient Greece and Egypt were able to self govern themselves and created a mutually beneficial relationship with it's constituents to the point of producing some of the greatest minds of the time (more so with Greece).
Here appears to be a case of it's use in the field of AI... Scary, but supposedly it's intent is for knowledge acquisition.
Update While I do like Bartek M's point of view, I have already come to realize this about neural networks. Yes, they are just 1's and 0's, and a static representation of a model that produces it's own approximated function estimation of some problem.
However, imagine a universal approximation machine (this is not insanse. Imagine calculus and integrals, yet each neuron is set to a specific input of a domain problem. ANN starts it's error checking based on input/output, and it applies a Genetic Algorithm of evolution.
Something that has the ability to reproduce 1000 times over while applying a minor mutation, pick the best performing 2 copies, and then following successive generations are produced. Somethings humans can't do internally.
I've just described a genetic algorithm of evolution. Some could argue it is also "learning". Only problem is... memories. However, databases serving as input solve the issue with memory. The ANN can serve as the brain or thought processor. In essence I think there is a formula for self awareness and consciousness at the bottom of all this.
Apply this to a social problem, like figure out how a society would operate, and you can function approximate down to a lot of levels.