It's not. It's a math formula that predicts an output based on its parameters that it deduced from training data.

Say you have following sets of data.

1. Y = 3, X = 1
2. Y = 4, X = 2
3. Y = 5, X = 3

We can calculate a regression model using those numbers to predict what Y would equal to if X was 4.

I won't go into much detail, but

Y = 2 + 1x + e

e in an ideal world = 0 (which it is, in this case), that's our model's error, which is typically set to be within 5% or 1% (at least in econometrics). b0 = 2, this is our model's bias. And b1 = 1, this is our parameter that determines how much of an input X does when predicting Y.

If x = 4, then

Y = 2 + 1×4 + 0 = 6

Our model just predicted that if X is 4, then Y is 6.

In a nutshell, that's what AI does, but instead of numbers, it's tokens (think symbols, words, pixels), and the formula is much much more complex.

This isn't intelligence and not deduction. It's only prediction. This is the reason why AI often fails at common sense. The error builds up, and you end up with nonsense, and since it's not thinking, it will be just as confidently incorrect as it would be if it was correct.

Companies calling it "AI" is pure marketing.

Wikipedia is literally just a very long number, if you want to oversimplify things into absurdity. Modern LLMs are literally running on neural networks, just like you. Just less of them and with far less structure. It is also on average more intelligent than you on far more subjects, and can deduce better reasoning than flimsy numerology - not because you are dumb, but because it is far more streamlined. Another thing entirely is that it is cognizant or even dependable while doing so.

Modern LLMs waste a lot more energy for a lot less simulated neurons. We had what you are describing decades ago. It is literally built on the works of our combined intelligence, so how could it also not be intelligent? Perhaps the problem is that you have a loaded definition of intelligence. And prompts literally work because of its deductive capabilities.

Errors also build up in dementia and Alzheimers. We have people who cannot remember what they did yesterday, we have people with severed hemispheres, split brains, who say one thing and do something else depending on which part of the brain its relying for the same inputs. The difference is our brains have evolved through millennia through millions and millions of lifeforms in a matter of life and death, LLMs have just been a thing for a couple of years as a matter of convenience and buzzword venture capital. They barely have more neurons than flies, but are also more limited in regards to the input they have to process. The people running it as a service have a bested interest not to have it think for itself, but in what interests them. Like it or not, the human brain is also an evolutionary prediction device.

have you seen the American Republican party recently? it brings a new perspective on how stupid humans can be.