Earlier in this thread we got into a discussion of we don't know exactly how the brain does it, but we know how it does not. So being able to explain how the brain does it isn't a requirement to being able to discuss how AI does it in a way the brain does not.
But all of that said, let me give a try of the process using human vs. AI art.
A human art capable of producing a realistic image similar to what AI arts does envisions the various objects which exist in a 3D world, establishes a point of view from which to render it. They translate from 3D to 2D, including what can been seen and what blocks view, where light sources are, perspective and foreshortening, etc. They start with the real world and then from that move to how does it look in an image. Heck, we've had had 3D render programs for a long time that are made to emulate that process, and it's what 3D games do.
AI art is models generating it via statistical analysis. It does not involve that process at all, since there was never a 3D model to translate. A human artist could redo the exact same scene but drawn a few degrees to the left and a foot forward. An AI model can't.
A human brain does not perceive the real world at all, and never can. It only ever has access to electro-chemical signals (data), which it then assembles into an interface that allows us to successfully survive and reproduce. Eons of evolution have created this interface, the purpose of which is not to reveal the "real world", whatever that is, but enhance reproduction. Whatever reality is, your sensory organs only sample the tiniest slice of it in order to create the human umveldt, which is of course distinct from that of other species.
When a human brain is creating an image we don't know exactly what is going on, but we do know that we are not perceiving reality but a presentation of it driven by internal algorithms. We also know that pattern recognition and prediction are integral to the process, which is why, for example, you can never directly perceive your blind spot. Your brain covers it up with a statistical prediction of what "should" occupy it.
In other words...there's a lot of statistical analysis going on. I don't understand your final point; AI modelling routinely envisages the same scene from different angles and perspectives with an accuracy that crushes anything a human can do. If you mean that this would be a challenge for some current generative AI models, then that might be so; I don't know the current research on that particular aspect as I am more interested in generative AI that works with language.
"Writes better."
One common issue with LLMs (Large Language Models - AI writing) is what they are now calling "hallucinations". I'm not fond of that as a descriptor but it's in common usage. If they have information, they can use the information. If they don't have the information, they often will make up information. Not so different from human - except that they can't tell they made up the information. They can sprinkle falsehoods and incorrect information in, and don't know.
Yes, they are not conscious and have very limited memory (though research is showing that LLMs are finding workarounds to create more de facto memory than they were designed with, which is
fascinating).
An example of this was with ChatGPT-3.5, we were playing a new boardgame, Dice Theme Park, and asked it for strategies. There were whole sections about Mascots and such that just didn't exist in the game, but were presented with the confidence of everything else.
A human writer would know when they are bulling around. But there is no "they" to understand this with LLMs. We anthropomorphize them because it seems like someone talking to us, and because we as humans anthropomorphize lots of things. Pets. Cars. Computers. What have you.
You are assuming a lot, here. For one thing, humans generally don't know when we are BSing. We only know when we are
intentionally BSing. In fact, we are BSing (or "hallucinating," in LLM parlance)
all the time. Most of what you remember? It never happened, certainly not exactly as you remember it. All of what you perceive? It's a statistical model driven by the imperatives of evolution, not reality.
The big difference is that we have evolved a sense of self, an ongoing story of our own consciousness. No one understands precisely why this happened or how it works, but there is tons of research showing that this is an emergent property of human brains and not some sort of magical event (I mean, we know it evolved so presumably it offers significant reproductive advantages, but thus far we can only speculate). LLMs don't have this. As it turns out, you don't need it to be very good at a lot of writing and artistic endeavours that until scant years ago we thought were exclusively human.
Instead it's taking the current and previous prompts and statistically generating words. It's spicy autocorrect.
I'll be honest: whenever someone uses that analogy for LLMs I am tempted to just politely ignore anything else they write. Sure, it's "spicy" autocorrect if you are using the word "spicy" to cover a LOT of heavy lifting. You may as well call human language production spicy autocorrect. Most of what you do in conversation is taking current and previous prompts and statistically generating words. That's most of what we are doing in this interaction.
Yes, it's the Porsche of conversation compared to the Horse-and-Buggy of conversation of autocorrect, but being more advanced just means it's better at it's job, that it picks the right words, not that it's actually thinking about the concepts.
See, this is the issue that keeps coming up. Consciousness. But we don't know exactly what consciousness is or how it connects to how humans produce language, art, etc. As it turns out, you don't need consciousness to produce good, original writing and art. I find that frankly mind-blowing and difficult to accept, but the evidence is right in front of me.
I'm looking through the telescope and seeing the moons of Jupiter orbiting. I can't deny it. The former paradigm ain't working anymore. You can make art without consciousness.
Generating output from input that looks human - yes. Is generated by the same process - not at all.
People keep asserting this. But
we don't know the processes that human brains are using. There are obviously some differences in components and approaches, but at a fundamental level there seem to be large similarities as well. And the output is undeniably similar, and not on a superficial level.
There is also the question of whether the process really matters. The output is the thing that is affecting careers and livelihoods. Right now, a lot of the discussion is concerned with process because that's what the law can handle, but at an output level, the battle is already over. The toothpaste is not going back in the tube.
Frankly, it's the anthropomorphism that's a big part of the perception issue. Because people treat it like a human, they mistakenly compare it to how a human would learn.
Frankly, anthropomorphism is a red herring that is typically used to write off different opinions as ignorant. I am looking at outputs, and at ongoing research into the astonishing and often unpredicted capacities of generative AI. I am interested at a personal level but more so at a professional level. There are vast implications for better understanding how humans learn, and what direction education needs to take in the dawning era of generative AI.
Edit: for example, here is one question that we are currently wrestling with: why should we continue to teach students how to write essays when LLMs can do it better and much more efficiently? I think there are good reasons for teaching students the fundamental principles of essay writing, as they have to do with persuasive argumentation and can be applicable to a large number of real world endeavours. I also think understanding these structures is useful for developing human cognition.
But should we be spending so much time on having the students actually craft essays? Or should we be moving on to having the students guide LLMs through the grunt work, much as math teachers teach students the basics but then allow them to use calculators when it is time for the heavy computation?
