This statement implies that humans can be more or less special "actually", as if it were a matter of fact, of objective reality.

Well yes, yes it does. Even if "specialness" is defined purely within human neurology doesn't mean you can't apply it's criteria to parts of reality and be objectively right or wrong about the result - just like, say, numbers.

Now, you could argue that humans vary with regards to how "special" humanity is to them, I suppose ... but in practice we seem to have a common cause, generally. Alternately, you could complain that paperclippers disagree about our "specialness" (or rather mean something different by the term, since their specialness algorithm returns high values for paperclips and low ones for humans and rocks), and is therefore insufficiently objective, but ...