Hubert Dreyfus among others have risen to prominence in philosophy of AI by emphasising (among other things) the directive primacy of unconscious or automatic, habit-driven processes. Dreyfus came to recognise this as a problem for AI developers working with formal symbol manipulation matrices through his readings in the phenomenology tradition, namely Husserl and his insights on the directed nature of consciousness.
Influenced heavily by the same tradition, Sartre famously characterised consciousness as a 'nothingness', an 'absence' or a 'Pierre-who-is-never-there', an entity which identity simply does not hold to in the same way as with objects in the world. A waiter cannot be a waiter in the same way a tree is a tree, to think so is to fall into bad faith, to 'be not what one is and to be what one is not' .. Derrida describes this beautifully in saying Sartre's formulation is a 're-apropriation of negativity'.
Does Sartre's conception of consciousness imply the impossibility of a conscious computer?