This seems very similar to the research program led by the late Patrick Henry Winton: https://groups.csail.mit.edu/genesis/index.html

Besides, I wish that causality had been mentioned more than once in passing. Due to the existence of the ladder of causality, many important queries cannot be answered by mere observation, or even by intervention; such queries require counterfactual reasoning, and structural causal models generalize because they describe something that is very invariant in the world.