Artificial Intelligence is a flawed concept
I was just reading through wikipedia and read (very briefly and cursorily) some articles on neuroscience.
From them I gleaned that the brain can be divided into
1. Neocortex: intelligence - basically logic
2. Visceral/Reptilian brain: emotions, basic sensations
According to neuroscience the neocortex is a later development i.e. we share our visceral brains with all animals but are different from them in that we have a neocortex.
So, I'm not wrong in asserting that logic is the final evolutionary endpoint as far as the brain is concerned. We've already replicated that in computers.
To then think of creating actual humans (artificial intelligence AI) would be to try and mimick the reptilian parts of our brain and that I regard as a step backwards instead of forwards as most AI researchers assume.
To make it even more obvious that we've already achieved the ultimate goal of AI we need only realize that we can already construct robots that mimick the reptilian brain that, to me, is a very basic input-output device with every possible combination of behavior being hardwired and, well, subconscious.
So, humans are not AI researchers need to create. What they actually need to do is create the perfect logical machine which appears to be already accomplished.
AI development in its current form is flawed.
Your views...
From them I gleaned that the brain can be divided into
1. Neocortex: intelligence - basically logic
2. Visceral/Reptilian brain: emotions, basic sensations
According to neuroscience the neocortex is a later development i.e. we share our visceral brains with all animals but are different from them in that we have a neocortex.
So, I'm not wrong in asserting that logic is the final evolutionary endpoint as far as the brain is concerned. We've already replicated that in computers.
To then think of creating actual humans (artificial intelligence AI) would be to try and mimick the reptilian parts of our brain and that I regard as a step backwards instead of forwards as most AI researchers assume.
To make it even more obvious that we've already achieved the ultimate goal of AI we need only realize that we can already construct robots that mimick the reptilian brain that, to me, is a very basic input-output device with every possible combination of behavior being hardwired and, well, subconscious.
So, humans are not AI researchers need to create. What they actually need to do is create the perfect logical machine which appears to be already accomplished.
AI development in its current form is flawed.
Your views...
Comments (9)
No.
No.
No.
No.
Humans are the only example of large general intelligence there is, so Ai researchers work from there.
Animals have had limbic systems for a very long time. Birds whose lineage is quite ancient, have emotions--maybe not as complicated as ours, but emotions none the less. Our limbic system is tied into structures like the pre-frontal cortex. It's a critical connection: It's what enables us to learn right from wrong, feel guilt, and avoid behavior that makes us feel terrible. If those connections aren't working, we behave psychopathically.
Our neocortex is nicely complicated, but animals have had the capacity to do a little thinking for quite a long time. Animals that coordinate hunting, for instance, have to 'think' about what is going on as they stalk and chase down prey -- where they, as an individual, fit into the hunt. Coordination, in other words. It may not be Aristotle, but it's thinking.
Dogs manipulate us. They have just enough intelligence to figure out how to get us to do what they want us to do. Quite often we adjust our behavior to satisfy the dog's wishes. We like observing their naked thinking at work. It's amusing.
For us, intelligence is a combination of flesh, emotion, memory, perception, and thinking all rolled up together. The best IBM computer both has and lacks some aspects of what we define in ourselves as intelligence. It has some limited perception (input devices), it has a memory, and it has logical processors (which in themselves do not constitute a capacity to think). It lacks a body (flesh, blood), emotion, and the wide scope of our perceptions and thinking. These are differences in kind, not just in quantity. The IBM had no desire to learn anything. It had no desire to engage in a debate. It didn't know it was engaging in a debate. What it was doing was executing commands in a very complex human-authored program and using brute force to assemble and organize information.
I'm not knocking the IBM and engineer's achievement. It's pretty impressive. However, the computer didn't really achieve anything.
IF one day a computer voluntarily experienced a desire to learn human language and culture, learn about the world, learn how to move about and manipulate the world, and find it's own place in the world, expressed likes and dislikes and acted accordingly, and did all this on its own, I'd call that REAL artificial intelligence.
What we have so far are machines that run human authored programs that imitate certain aspects of behavior.
Well, if the neuroscience I've read is true then we (humans) are at the tip of the spear of brain evolution.
What does that mean?
Neocortex (logical thinking) and not the primitive reptilian brain that forms the "rest" of our psyche (emotional and less rational).
So, if AI is supposed to be human-like it would mean that we have to somehow code for the reptilian part of the human brain. I'm not saying it can't be done. I think it is easier for higher intelligence to mimic lower intelligence than the other way round. However, this would not be a step forward as assumed by AI researchers. It would actually count as regression.
Did you mean logistic regression?
Your speech shows that you really don't know what you're on about.
https://machinelearningmastery.com/logistic-regression-for-machine-learning/
However, computers are not currently infallible in their algorithms. Without a consciousness, they are unable to develop solutions beyond the bounds of preprogrammed instructions. Certain mathematical inputs will return an error; calculators are unable to innovate by using the correct branch of abstract mathematics. It is inevitable that any program which is currently created will contain fault; even the most prodigious human solvers are not infallible.
To create a mind without instinct and affect might have consequences unforeseen. While I know relatively little of neurobiology or of psychology, I am aware that the human mind is a layered construct. Our childhood memories, our keen wounds and triumphs, and our fixations culminate into the adult mind. Our affect and instincts inform our logical conclusions. Therefore, it is necessary to perform careful studies of multiple fields to construct our superior intelligence.