Everything must be clearly, and at the very least : in important ways tied to that which is real empiricism: DIRECT observation of most-relevant and major OVERT behavior patterns (including the observable concrete shifts in behavior patterns, i.e. further patterning).
This empiricism is true of all science and would/will be in Psychology, IFF Psychology is to be a science -- which it clearly is NOT yet. (Learn the nature of other science WHICH PROGRESSES -- and really IS science -- to learn the Reality [(truth)] of this.)
"NEVER 'Models first ...' "
P.S. I have tried to provide a full outline of such a science of Psychology (and because it is truly understandable [(essentially: concrete)], it is good for AI too). If you have not been motivated to "check me out", you lack good critical abilities (and are hopelessly deluded, if you think you have a SCIENCE -- this is fact/Reality, not an opinion).
It is my hope that you will read both of my major writings:
Not only https://www.researchgate.net/publication/286920820_A_Human_Ethogram_Its_Scientific_Acceptability_and_Importance_now_NEW_because_new_technology_allows_investigation_of_the_hypotheses_an_early_MUST_READ
but also the much, much more recent:
https://www.researchgate.net/publication/329428629_Essentially_all_Recent_Essays_on_Ethogram_Theory
I an seriously trying to have a perspective and approach which has all the features and dimensions a science theory should .