Yeah I guess you could make the case for it being the other way around. Although, I think that a lot of ML concepts have groundings in optimization, probability, and a bunch of other mathematical concepts as opposed to concepts that wanted to replicate those psychological effects.