The thing with AI generation for art is that it needs guiding human input, and in order to make the most effective final result you need to instruct it more and more specifically until you are essentially just "painting" yourself, since that is what painting is fundamentally - complete control over every aspect of a picture. The same would apply to any ai generated product. I imagine it will help expedite more tedious parts however (aka "just a tool").
Aesthetic inquiry should start with foods. I think our experience with it is fundamentally the same as our consumption of other "arts" but it also provides us with more traceable 'cause and effect' which is more easily formed into conclusions given that the source does not travel far to become result. * I imagine a substantial portion of our seemingly more abstract visual aesthetic preferences derive from our biologically programmed repulsion to diseased organic material. Humans (to varying individual extents) are adept at pattern-matching, abstraction etc and thus we take these characteristics associated with displeasure and the converse with pleasure and they are applied to ornamentation in general. Is our preference for symmetrical faces thus reflected in the first symmetrical arrowheads and blades of tools then? Are the quantitative relations so strongly imprinted that we have an intuitive aural sense for them as well (appreciating patterns in music)? Or does that come from another organ?
Narrow vertical/natural learning: learning in whatever order things happen to be presented to you. Very intuitive and often operates on implicit lessons rather than explicitly noting mechanics/interactions. Lots of learning by mimesis.
Systematic learning: More care is taken to foundations and definitions. Important axioms are contextualized as useful for the core lessons. However, ideas will often just be taken on premise to avoid long tangents that don't improve someone's understanding of the subject in at least the mid-term (as opposed to long-term).
Horizontal/highly systematic learning: done only for the sake of learning something at leisure.
To no one's surprise except my own 3 years ago, specific domain expertise (some field of math, let's say) does not generalize even into capacity for broader critical thinking necessarily. * I've understated the influence that prior recognition with a pattern has. To get somewhere further we have to have flattened some patterns to shorthand so a larger compound concept may be understood. Thus it's the facility for flattening patterns that allows us to understand more..? I guess that's just pattern recognition. The more something can be relegated to an existing "line" that already has an appropriate response primed the more mental space is available for other considerations. * quality of reasoning matters more than the conclusion