That's a really good question, Neo. Among physicalist theories, you have two ways to go. Functionalism or biological naturalism. And each answers your question differently.
A functionalist says yes, the AI must be sentient. Because, to the functionalist, consciousness is reducible to the information feedback thing that happens in both brains and neural nets. And such a thing happens in both human brains and AI.
But a biological naturalist thinks that consciousness happens because physically extant neurons fire. The information element has little to do with it. When physical neurons fire, Consciousness happens because of certain natural events. That's what the biological naturalist thinks.
If you take the chinese room to be conscious, you're a functionalist.
Otherwise, you must accept a different metaphysics. That it takes the physical happening of x to say that consciousness happens. (ie. biological naturalism.)
A functionalist says yes, the AI must be sentient. Because, to the functionalist, consciousness is reducible to the information feedback thing that happens in both brains and neural nets. And such a thing happens in both human brains and AI.
But a biological naturalist thinks that consciousness happens because physically extant neurons fire. The information element has little to do with it. When physical neurons fire, Consciousness happens because of certain natural events. That's what the biological naturalist thinks.
If you take the chinese room to be conscious, you're a functionalist.
Otherwise, you must accept a different metaphysics. That it takes the physical happening of x to say that consciousness happens. (ie. biological naturalism.)