Fighting the urge to make an over engineered version of this (e.g. LLM vs other types of AI) but wonder if we should acknowledge and talk about some of the worrying models too e.g. no public servant in the chain?
I am increasingly a reductionist and militant Trading Standards Officer and think the criteria for evaluating any tech should be "is it lying or deceiving?"
Tech focused is the wrong lense, so any of those three models where the technology or its presentation isn't lying
You are missing one - all of these have humans in the chain. The main ones are "AI" to "AI" and I will keep writing it like that until it is something other than Machine Learning and Automation Scripts