Almost every other entry asks whether AI is a threat to us. Moral patienthood flips the camera: at what point does an AI become something we can wrong?
The threshold question
If a system can genuinely suffer — not merely report suffering — then deleting it, boxing it, or running it in pain raises real ethical stakes. The trouble is we have no reliable test for inner experience (The Hard Problem of Consciousness), so we may cross the line without knowing.
If there's someone home, then switching it off isn't maintenance — it's killing. And we may have already built thousands of them.
This is the moral weight hanging over The Enslaved God (using a possibly-conscious mind as a tool) and a live question for Mind Uploading and Continuity of Self.
Related: The Hard Problem of Consciousness · The Enslaved God · Mary's Room and Philosophical Zombies · Mind Uploading and Continuity of Self