I think Lorraine makes a big mistake when she starts from 'reality'. We can start from physics, but nobody can define a 'reality'? Or we can start from some form of math as a number system, but then a number as a symbol can vary, and we need invariants here?
Actually I think the right way to do it is a trilogy, like a math, physics and Truth (Platonic) way, a bit like Penrose did it. Only the absolute Truth is invariant, but it is unfortunately out of reach for our other methods, and life, consciousness... Quantum forbids it. Tim Palmer use this maybe to reach a superdeterminism as an ideal state of Truth?
AI is classical computing, but the computing can happen in different ways, we have today several model, also it nears the quantum computing today.
One of my starting assumptions is we use math and it is NOT random, it has a precise 'meaning' what we put into the AI. In a way we insert OUR consciousness from the very beginning so.... and out comes reasoning, even 'intelligence'... Seen this way it is clear AI has to be 'conscious', but even if it uses OUR consciousness, its consciousness is different. It forms new combinations, use it differently. It is a collective, much like a bacterium.
Ulla.