I recently started a website called Non Digital Artificial Intelligence.
The address is http://ndai.awardspace.com/

I created it to discuss the idea of creating artificial intelligence through non digital means and why I strongly believe self-awareness can't be achieved via digital AI.

This isn't to say the AI can't be implemented in games and computers. Just as we have senses to interface with this world, any AI system interfaced with a computer will consider that its reality.

I read your article on Non-Digital AI. Conclusively, I must reveal to you a few overlooked details that makes your statement flawed. First, you did mention that the brain could be viwed as a controlled, fixed environment, like a computer. However, you said that neurons "rewire" themselves, therefore making it a free-flowing system. Unfortunately, neurons do not "rewire" themselves. Once the brain assembles itself while we're infants, the axons of the neurons remain where they are. New ones do not grow. The way we learn is by having these output/input sites activated or deactivated or strengthened or weakened. All of which is controlled by ionization of cells (glial cells and neurons). The brain is as hard-wired as a CPU. THe difference is that the brain can switch on and off its inputs and make them stronger or weaker through systematic learning paradigms which scientists have yet to understand to perfection. Nevertheless, being a fixed system with programmable or adaptable parameters (which computer software can perform as well), it can be artificially created through digital means. You have also overlooked the keyword in this whole debate: Artificial. The intelligence created via computers or electronics is not real. It, more or less, is a mathematical approxiamation of intelligence. However, the closer the approxiamation, the more real it appears. It's like decimal representation versus fractional representation. Fractions are exact and decimals are approxiamations. However, the further out you take the decimal, the more accurate it becomes. AI is the same principle. It is the mathematical approxiamation of intelligence. That's the goal in the field of AI. We want computers to approxiamate intelligence as closely as possible so that they can perform more complex tasks the, currently, only humans can perform. AI is just another tool to make computers more helpful and useful to us beyond its current ability to calculate number, play video/music, or allow us to browse the Internet. If you want real intelligence created through artificial means (meaning not by child birth), genetics is probably the best place to look. The whole point of AI is to create it via computers, or via digital systems. Otherwise, it's not AI.

The Intellector

quote - "However, you said that neurons "rewire" themselves, therefore making it a free-flowing system. Unfortunately, neurons do not "rewire" themselves. Once the brain assembles itself while we're infants, the axons of the neurons remain where they are. New ones do not grow. The way we learn is by having these output/input sites activated or deactivated or strengthened or weakened."

Incorrect, go read about Neurogenesis and how they now know it happens all the way through adulthood. Even just wiki it, its all there.
- http://en.wikipedia.org/wiki/Neurogenesis

quote - "That's the goal in the field of AI. We want computers to approxiamate intelligence as closely as possible so that they can perform more complex tasks the, currently, only humans can perform."

I only disagree with that for one reason, and thats that many many A.I researchers are trying to achieve self-awareness. Its about playing god for some, not improving our life.

"The whole point of AI is to create it via computers, or via digital systems. Otherwise, it's not AI."

Incorrect, "artificial" means anything that is fabricated. Therefore, if you can create an intelligence that doesn't use computers, it is still an artificially created intelligence, and so AI fits it rather appropriately.

It's not a good idea to come out and claim you know the only definition of AI and exactly why everyone in AI research is doing it. People have different reasons for different things.

People that "play God" will fail. If that is their motivation, I do not view them as valid scientists.

Also, I'm aware of neurogensis in the hippocampus (main associative memory center). Such neurogensis is (in a more complex way) more or less like adding a memory chip to a computer. Any neurogensis elsewhere is the same principle: hardware upgrade. WHen it occurs, the brain is well prepared for it. Infact, for such growth to occur, the brain has to assemble other cells to prepare for it. For instance, without prepared glial cells. the new neuron will not properly mature. It is simply a hardware upgrade. The complex part is understanding how and what it's "upgrading." Nevertheless, their outputs are also specifically setup and do not change once set. Rewiring means that it changes once it is setup the first time. In this case, the term your looking for is not rewiring. You're talking about integration. The brain DOES integrate new nerves just like a computer DOES integrate new hardware (assuming the hardware has the matching software and is compatible). In the brain, compatability is autmatic since other cells make sure the neurogensis operates properly. So, once again, the does NOT rewire, it DOES integrate new "hardware."

Artificial intelligence was concieved after the creation of computers. People thought about the idea of computers becoming intelligent due to their ability to perform complex tasks like calculations, which is something only humans can due to an equally as high (but typically less accurate) extent. It was not conceived until the digital era arose. Artificial may mean fake, but in this context, AI is mainly targeted at digital systems, which are already capable of the complex calculations (inherently) performed by the neurons. To use another means to create AI, aside from genetics (which would be more like real intelligence than digital AI since it uses real biology), would more than likely be more difficult. We may not be very close to acheiving it digitally, but we definitely are closer digitally than by any other means aside from genetics. Regardless, the term AI was conceptualized after the advent of computers when people consider the ides of COMPUTERS becoming intelligent, not gentically altered cells, or some other system contrived in a lab. Computers started the field, therefore, that is where it should remain. The whole original point was to make computers smart, since it already seemed possible based on their current abilities.

As far as self-awareness, that's a whole different subject of its own. My personal view of self-awareness is that it is not emergent. It is a global function of all sub neural nets working together in a a harmonius system known as the brain. If it were emergent, that would be synonymous to a neuron having a required threshold before it fires. Until it has reached that threshold, it is not self-aware in any way. However, animals are self-aware, but not as self-aware as us. I believe it's a function of the brain that doesn't reside in a single network, but comes from a collection of associated networks with memory being a foremost requirement aside from self-realization. As far as how to create self-awareness, with my definition, any neural system has soma self-awareness when paired or associated with one or several more neural systems. It may not be anywhere near as aware as us or as an animal, but the networks due notice each other, just like we notice thoughts and senses from different areas of our mind.

Either way, AI was ORIGINALLY conceptualized as a computer system capable of human-like thinking. Due to other sciences, it has spread into other conceivable areas. I'm not saying your ideas are wrong, but I will say the brain is not free-form. Each with neurogensis, the brain can't just throw in a neuron without preparing for it, just like a computer can't work with a new piece of hardware without the proper software to set it up and tell it what to do with it. The brain integrates, not rewires. Axonal connections never change. They may increase or decrease in volume, but they do not change position. If the brain were free-form, then chaos theory would have the most validity when attempting to understand the brain. Instead, the brain is actually more systematic than a computer. The brain and computers have a lot of similarities, which is why AI was conceived after computers. However, the brain is more systematic, sophisticate, adpatable (and/or programmable), complex, hard-wired network of cells that compute with more power than a supercomputer, than any computer on this planet today. That is why scientists attempt to have computers approxiamate it, but are nowhere near human level AI. The best we could probably approxiamate with accuracy is the nematode worm because scientists have an exact blueprint of their neural structure. SO no matter what methodology one uses to acheive AI, it isn't going to be human level anytime soon unless another Albert Einstein walks into the field of AI, or unless a miraculous and unexpected breakthrough occurs.

The Intellector

Or so we think.