Interestingly the criticism of LLM based AI's is that they are only prediction engines, attempting to predict the next word or phrase.

This understanding is correct.

The lack of understanding, however, is that is what we, as humans,

I left that sentence unfinished to prove my point.