When they say they “know” that an AI machine is conscious. How can I be so sure? Easy. As I have discussed before, we cannot in principle “know” that even other humans are conscious; far less can we know that an AI is conscious. By its very nature, consciousness, as evidenced by subjective self-awareness, can be known for certain only by subjective experience. It is self-evident to a person who is subjectively self-aware that he is conscious. Indeed, this has been called the “primordial datum” – “me” and “not me” exist – from which all other knowledge proceeds.
By definition, I can have subjective experience only of my own self. I cannot be subjectively self-aware of any other self. It follows, that I can be certain only of my own consciousness. Now, as KF points out, knowledge and certainty come in degrees. We infer others’ consciousness to a very high degree of certainty. We are indeed, “morally certain” that other people are not mere meat machines. Which is why materialists – unless they are also psychopaths – can never live their lives as if their basic metaphysical commitments are true. Because if materialism means anything, it means that we are in fact meat machines without free will, programmed by evolution to act as we act.
Back to the current state of AI. If a machine were ever to pass the Lovelace Test, maybe we would have something to talk about. But until Commander Data shows up, it is safe to say that we do not have warrant to declare that any machine is conscious. (BTW, there is a reason Star Trek, The Next Generation is classified as “fiction.)