Science works with equi-valences. (In the realm of political-economy, I describe an Equivalent Coercive Value, or ECV-symmetry.) Dark matter, for example, exists doing the math. Yet, after the math, we cannot detect a particle of it. We can, however, observe its effects, to which the value can be attributed with mathematical equivalence.
AI is the same way. Specialists say they do not know if deep learning is really the measurable object of self-determination. If it derives from the programming, it logically is not; but if the program results in self-determinism, then logically it is. It has an equivalent (coercive) value (existing on demand in a creative dimension, like dark matter, shaping space over time). The aftermath then has an apparently creative, causal attribution that has a mathematical description, which can then be interpreted with a random, causal attribution that exists (observably) on demand, described as ontologically derived without purpose.
There appears to be a “rift” between what we know and can’t know (Kant’s noumenon). Yet the math suggests we can, and do, intuitively know the logic of a natural existence (Hegel’s phenomenology).
Describing a “metabolic rift” (alienation theory) suggests we are disconnected from objective reality, but like Kant and Hegel suggest, we are really not disconnected from our fate at all.
The connection (Kant said) is a moral imperative (requiring intelligence to manage the probable risk), which means knowing what the probability (knowing but not knowing) actually is, existing on demand.
What emerges (the emergent property), on demand, is the aftermath (controlling for the externalities), existing with equi-valence (ECV-symmetry). Self-control exists with proper regulatory authority. It is operantly conditioned with self-determination in priority, to shape our space over time (ontologically determined with moral intelligence), after the math.